repo_id
stringlengths
21
96
file_path
stringlengths
31
155
content
stringlengths
1
92.9M
__index_level_0__
int64
0
0
rapidsai_public_repos/raft/python/pylibraft/pylibraft
rapidsai_public_repos/raft/python/pylibraft/pylibraft/common/mdspan.pxd
# # Copyright (c) 2022-2023, NVIDIA CORPORATION. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # # cython: profile=False # distutils: language = c++ # cython: embedsignature = True # cython: language_level = 3 from libc.stdint cimport int8_t, int64_t, uint8_t, uint32_t from libcpp.string cimport string from pylibraft.common.cpp.mdspan cimport ( device_matrix_view, host_matrix_view, row_major, ) from pylibraft.common.handle cimport device_resources from pylibraft.common.optional cimport make_optional, optional # Cython doesn't like `const float` inside template parameters # hack around this with using typedefs ctypedef const float const_float ctypedef const int8_t const_int8_t ctypedef const uint8_t const_uint8_t cdef device_matrix_view[float, int64_t, row_major] get_dmv_float( array, check_shape) except * cdef device_matrix_view[uint8_t, int64_t, row_major] get_dmv_uint8( array, check_shape) except * cdef device_matrix_view[int8_t, int64_t, row_major] get_dmv_int8( array, check_shape) except * cdef device_matrix_view[int64_t, int64_t, row_major] get_dmv_int64( array, check_shape) except * cdef optional[device_matrix_view[int64_t, int64_t, row_major]] make_optional_view_int64( # noqa: E501 device_matrix_view[int64_t, int64_t, row_major]& dmv) except * cdef device_matrix_view[uint32_t, int64_t, row_major] get_dmv_uint32( array, check_shape) except * cdef device_matrix_view[const_float, int64_t, row_major] get_const_dmv_float( array, check_shape) except * cdef device_matrix_view[const_uint8_t, int64_t, row_major] get_const_dmv_uint8( array, check_shape) except * cdef device_matrix_view[const_int8_t, int64_t, row_major] get_const_dmv_int8( array, check_shape) except * cdef host_matrix_view[float, int64_t, row_major] get_hmv_float( array, check_shape) except * cdef host_matrix_view[uint8_t, int64_t, row_major] get_hmv_uint8( array, check_shape) except * cdef host_matrix_view[int8_t, int64_t, row_major] get_hmv_int8( array, check_shape) except * cdef host_matrix_view[int64_t, int64_t, row_major] get_hmv_int64( array, check_shape) except * cdef host_matrix_view[uint32_t, int64_t, row_major] get_hmv_uint32( array, check_shape) except * cdef host_matrix_view[const_float, int64_t, row_major] get_const_hmv_float( array, check_shape) except * cdef host_matrix_view[const_uint8_t, int64_t, row_major] get_const_hmv_uint8( array, check_shape) except * cdef host_matrix_view[const_int8_t, int64_t, row_major] get_const_hmv_int8( array, check_shape) except *
0
rapidsai_public_repos/raft/python/pylibraft/pylibraft/common
rapidsai_public_repos/raft/python/pylibraft/pylibraft/common/cpp/optional.pxd
# # Copyright (c) 2022, NVIDIA CORPORATION. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # We're still using cython v0.29.x - which doesn't have std::optional # support. Include the minimal definition here as suggested by # https://github.com/cython/cython/issues/3293#issuecomment-1223058101 cdef extern from "<optional>" namespace "std" nogil: cdef cppclass optional[T]: optional() optional& operator=[U](U&)
0
rapidsai_public_repos/raft/python/pylibraft/pylibraft/common
rapidsai_public_repos/raft/python/pylibraft/pylibraft/common/cpp/mdspan.pxd
# # Copyright (c) 2022-2023, NVIDIA CORPORATION. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # # cython: profile=False # distutils: language = c++ # cython: embedsignature = True # cython: language_level = 3 from libc.stdint cimport int8_t, int64_t, uint8_t, uint64_t from libcpp.string cimport string from pylibraft.common.handle cimport device_resources cdef extern from "raft/thirdparty/mdspan/include/experimental/__p0009_bits/layout_stride.hpp" namespace "std::experimental": # noqa: E501 cdef cppclass layout_right: pass cdef cppclass layout_left: pass cdef extern from "raft/core/mdspan_types.hpp" \ namespace "raft": ctypedef layout_right row_major ctypedef layout_left col_major cdef cppclass matrix_extent[IndexType]: pass cdef extern from "raft/core/device_mdspan.hpp" namespace "raft" nogil: cdef cppclass device_vector_view[ElementType, IndexType]: pass cdef cppclass device_scalar_view[ElementType, IndexType]: pass cdef cppclass device_matrix_view[ElementType, IndexType, LayoutType]: pass cdef device_matrix_view[ElementType, IndexType, LayoutPolicy] \ make_device_matrix_view[ElementType, IndexType, LayoutPolicy]( ElementType* ptr, IndexType n_rows, IndexType n_cols) except + cdef device_vector_view[ElementType, IndexType] \ make_device_vector_view[ElementType, IndexType]( ElementType* ptr, IndexType n) except + cdef device_scalar_view[ElementType, IndexType] \ make_device_vector_view[ElementType, IndexType]( ElementType* ptr) except + cdef extern from "raft/core/host_mdspan.hpp" \ namespace "raft" nogil: cdef cppclass host_matrix_view[ElementType, IndexType, LayoutPolicy]: pass cdef cppclass host_vector_view[ElementType, IndexType]: pass cdef cppclass host_scalar_view[ElementType, IndexType]: pass cdef cppclass host_mdspan[ElementType, Extents, LayoutPolicy]: pass cdef host_matrix_view[ElementType, IndexType, LayoutPolicy] \ make_host_matrix_view[ElementType, IndexType, LayoutPolicy]( ElementType* ptr, IndexType n_rows, IndexType n_cols) except + cdef host_vector_view[ElementType, IndexType] \ make_host_vector_view[ElementType, IndexType]( ElementType* ptr, IndexType n) except + cdef host_scalar_view[ElementType, IndexType] \ make_host_scalar_view[ElementType, IndexType]( ElementType *ptr) except + cdef extern from "<sstream>" namespace "std" nogil: cdef cppclass ostringstream: ostringstream() except + string str() except + cdef extern from "<ostream>" namespace "std" nogil: cdef cppclass ostream: pass cdef extern from "raft/core/mdspan.hpp" namespace "raft" nogil: cdef cppclass dextents[IndentType, Rank]: pass cdef extern from "raft/core/serialize.hpp" namespace "raft" nogil: cdef void serialize_mdspan[ElementType, Extents, LayoutPolicy]( const device_resources& handle, ostream& os, const host_mdspan[ElementType, Extents, LayoutPolicy]& obj)
0
rapidsai_public_repos/raft/python/pylibraft/pylibraft
rapidsai_public_repos/raft/python/pylibraft/pylibraft/cluster/CMakeLists.txt
# ============================================================================= # Copyright (c) 2022-2023, NVIDIA CORPORATION. # # Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except # in compliance with the License. You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software distributed under the License # is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express # or implied. See the License for the specific language governing permissions and limitations under # the License. # ============================================================================= # Set the list of Cython files to build set(cython_sources kmeans.pyx) set(linked_libraries raft::compiled) # Build all of the Cython targets rapids_cython_create_modules( CXX SOURCE_FILES "${cython_sources}" LINKED_LIBRARIES "${linked_libraries}" ASSOCIATED_TARGETS raft MODULE_PREFIX cluster_ )
0
rapidsai_public_repos/raft/python/pylibraft/pylibraft
rapidsai_public_repos/raft/python/pylibraft/pylibraft/cluster/__init__.pxd
# Copyright (c) 2022, NVIDIA CORPORATION. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. #
0
rapidsai_public_repos/raft/python/pylibraft/pylibraft
rapidsai_public_repos/raft/python/pylibraft/pylibraft/cluster/__init__.py
# Copyright (c) 2022-2023, NVIDIA CORPORATION. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # from .kmeans import ( KMeansParams, cluster_cost, compute_new_centroids, fit, init_plus_plus, ) __all__ = [ "KMeansParams", "cluster_cost", "compute_new_centroids", "fit", "init_plus_plus", ]
0
rapidsai_public_repos/raft/python/pylibraft/pylibraft
rapidsai_public_repos/raft/python/pylibraft/pylibraft/cluster/kmeans.pyx
# # Copyright (c) 2022-2023, NVIDIA CORPORATION. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # # cython: profile=False # distutils: language = c++ # cython: embedsignature = True # cython: language_level = 3 import numpy as np from cython.operator cimport dereference as deref from libc.stdint cimport uintptr_t from libcpp cimport nullptr from collections import namedtuple from enum import IntEnum from pylibraft.common import Handle, cai_wrapper, device_ndarray from pylibraft.common.handle import auto_sync_handle from pylibraft.common.handle cimport device_resources from pylibraft.random.cpp.rng_state cimport RngState from pylibraft.common.input_validation import * from pylibraft.distance import DISTANCE_TYPES from pylibraft.cluster.cpp cimport kmeans as cpp_kmeans, kmeans_types from pylibraft.cluster.cpp.kmeans cimport ( cluster_cost as cpp_cluster_cost, init_plus_plus as cpp_init_plus_plus, update_centroids, ) from pylibraft.common.cpp.mdspan cimport * from pylibraft.common.cpp.optional cimport optional from pylibraft.common.handle cimport device_resources from pylibraft.common import auto_convert_output @auto_sync_handle @auto_convert_output def compute_new_centroids(X, centroids, labels, new_centroids, sample_weights=None, weight_per_cluster=None, handle=None): """ Compute new centroids given an input matrix and existing centroids Parameters ---------- X : Input CUDA array interface compliant matrix shape (m, k) centroids : Input CUDA array interface compliant matrix shape (n_clusters, k) labels : Input CUDA array interface compliant matrix shape (m, 1) new_centroids : Writable CUDA array interface compliant matrix shape (n_clusters, k) sample_weights : Optional input CUDA array interface compliant matrix shape (n_clusters, 1) default: None weight_per_cluster : Optional writable CUDA array interface compliant matrix shape (n_clusters, 1) default: None batch_samples : Optional integer specifying the batch size for X to compute distances in batches. default: m batch_centroids : Optional integer specifying the batch size for centroids to compute distances in batches. default: n_clusters {handle_docstring} Examples -------- >>> import cupy as cp >>> from pylibraft.common import Handle >>> from pylibraft.cluster.kmeans import compute_new_centroids >>> # A single RAFT handle can optionally be reused across >>> # pylibraft functions. >>> handle = Handle() >>> n_samples = 5000 >>> n_features = 50 >>> n_clusters = 3 >>> X = cp.random.random_sample((n_samples, n_features), ... dtype=cp.float32) >>> centroids = cp.random.random_sample((n_clusters, n_features), ... dtype=cp.float32) ... >>> labels = cp.random.randint(0, high=n_clusters, size=n_samples, ... dtype=cp.int32) >>> new_centroids = cp.empty((n_clusters, n_features), ... dtype=cp.float32) >>> compute_new_centroids( ... X, centroids, labels, new_centroids, handle=handle ... ) >>> # pylibraft functions are often asynchronous so the >>> # handle needs to be explicitly synchronized >>> handle.sync() """ x_cai = X.__cuda_array_interface__ centroids_cai = centroids.__cuda_array_interface__ new_centroids_cai = new_centroids.__cuda_array_interface__ labels_cai = labels.__cuda_array_interface__ m = x_cai["shape"][0] x_k = x_cai["shape"][1] n_clusters = centroids_cai["shape"][0] centroids_k = centroids_cai["shape"][1] new_centroids_k = centroids_cai["shape"][1] x_dt = np.dtype(x_cai["typestr"]) centroids_dt = np.dtype(centroids_cai["typestr"]) new_centroids_dt = np.dtype(new_centroids_cai["typestr"]) labels_dt = np.dtype(labels_cai["typestr"]) if not do_cols_match(X, centroids): raise ValueError("X and centroids must have same number of columns.") if not do_rows_match(X, labels): raise ValueError("X and labels must have same number of rows") x_ptr = <uintptr_t>x_cai["data"][0] centroids_ptr = <uintptr_t>centroids_cai["data"][0] new_centroids_ptr = <uintptr_t>new_centroids_cai["data"][0] labels_ptr = <uintptr_t>labels_cai["data"][0] if sample_weights is not None: sample_weights_cai = sample_weights.__cuda_array_interface__ sample_weights_ptr = <uintptr_t>sample_weights_cai["data"][0] sample_weights_dt = np.dtype(sample_weights_cai["typestr"]) else: sample_weights_ptr = <uintptr_t>nullptr if weight_per_cluster is not None: weight_per_cluster_cai = weight_per_cluster.__cuda_array_interface__ weight_per_cluster_ptr = <uintptr_t>weight_per_cluster_cai["data"][0] weight_per_cluster_dt = np.dtype(weight_per_cluster_cai["typestr"]) else: weight_per_cluster_ptr = <uintptr_t>nullptr handle = handle if handle is not None else Handle() cdef device_resources *h = <device_resources*><size_t>handle.getHandle() x_c_contiguous = is_c_contiguous(x_cai) centroids_c_contiguous = is_c_contiguous(centroids_cai) new_centroids_c_contiguous = is_c_contiguous(new_centroids_cai) if not x_c_contiguous or not centroids_c_contiguous \ or not new_centroids_c_contiguous: raise ValueError("Inputs must all be c contiguous") if not do_dtypes_match(X, centroids, new_centroids): raise ValueError("Inputs must all have the same dtypes " "(float32 or float64)") if x_dt == np.float32: update_centroids(deref(h), <float*> x_ptr, <int> m, <int> x_k, <int> n_clusters, <float*> sample_weights_ptr, <float*> centroids_ptr, <int*> labels_ptr, <float*> new_centroids_ptr, <float*> weight_per_cluster_ptr) elif x_dt == np.float64: update_centroids(deref(h), <double*> x_ptr, <int> m, <int> x_k, <int> n_clusters, <double*> sample_weights_ptr, <double*> centroids_ptr, <int*> labels_ptr, <double*> new_centroids_ptr, <double*> weight_per_cluster_ptr) else: raise ValueError("dtype %s not supported" % x_dt) @auto_sync_handle @auto_convert_output def init_plus_plus(X, n_clusters=None, seed=None, handle=None, centroids=None): """ Compute initial centroids using the "kmeans++" algorithm. Parameters ---------- X : Input CUDA array interface compliant matrix shape (m, k) n_clusters : Number of clusters to select seed : Controls the random sampling of centroids centroids : Optional writable CUDA array interface compliant matrix shape (n_clusters, k). Use instead of passing `n_clusters`. {handle_docstring} Examples -------- >>> import cupy as cp >>> from pylibraft.cluster.kmeans import init_plus_plus >>> n_samples = 5000 >>> n_features = 50 >>> n_clusters = 3 >>> X = cp.random.random_sample((n_samples, n_features), ... dtype=cp.float32) >>> centroids = init_plus_plus(X, n_clusters) """ if (n_clusters is not None and centroids is not None and n_clusters != centroids.shape[0]): msg = ("Parameters 'n_clusters' and 'centroids' " "are exclusive. Only pass one at a time.") raise RuntimeError(msg) cdef device_resources *h = <device_resources*><size_t>handle.getHandle() X_cai = cai_wrapper(X) X_cai.validate_shape_dtype(expected_dims=2) dtype = X_cai.dtype if centroids is not None: n_clusters = centroids.shape[0] else: centroids_shape = (n_clusters, X_cai.shape[1]) centroids = device_ndarray.empty(centroids_shape, dtype=dtype) centroids_cai = cai_wrapper(centroids) # Can't set attributes of KMeansParameters after creating it, so taking # a detour via a dict to collect the possible constructor arguments params_ = dict(n_clusters=n_clusters) if seed is not None: params_["seed"] = seed params = KMeansParams(**params_) if dtype == np.float64: cpp_init_plus_plus( deref(h), params.c_obj, make_device_matrix_view[double, int, row_major]( <double *><uintptr_t>X_cai.data, <int>X_cai.shape[0], <int>X_cai.shape[1]), make_device_matrix_view[double, int, row_major]( <double *><uintptr_t>centroids_cai.data, <int>centroids_cai.shape[0], <int>centroids_cai.shape[1]), ) elif dtype == np.float32: cpp_init_plus_plus( deref(h), params.c_obj, make_device_matrix_view[float, int, row_major]( <float *><uintptr_t>X_cai.data, <int>X_cai.shape[0], <int>X_cai.shape[1]), make_device_matrix_view[float, int, row_major]( <float *><uintptr_t>centroids_cai.data, <int>centroids_cai.shape[0], <int>centroids_cai.shape[1]), ) else: raise ValueError(f"Unhandled dtype ({dtype}) for X.") return centroids @auto_sync_handle @auto_convert_output def cluster_cost(X, centroids, handle=None): """ Compute cluster cost given an input matrix and existing centroids Parameters ---------- X : Input CUDA array interface compliant matrix shape (m, k) centroids : Input CUDA array interface compliant matrix shape (n_clusters, k) {handle_docstring} Examples -------- >>> import cupy as cp >>> from pylibraft.cluster.kmeans import cluster_cost >>> n_samples = 5000 >>> n_features = 50 >>> n_clusters = 3 >>> X = cp.random.random_sample((n_samples, n_features), ... dtype=cp.float32) >>> centroids = cp.random.random_sample((n_clusters, n_features), ... dtype=cp.float32) >>> inertia = cluster_cost(X, centroids) """ x_cai = X.__cuda_array_interface__ centroids_cai = centroids.__cuda_array_interface__ m = x_cai["shape"][0] x_k = x_cai["shape"][1] n_clusters = centroids_cai["shape"][0] centroids_k = centroids_cai["shape"][1] x_dt = np.dtype(x_cai["typestr"]) centroids_dt = np.dtype(centroids_cai["typestr"]) if not do_cols_match(X, centroids): raise ValueError("X and centroids must have same number of columns.") x_ptr = <uintptr_t>x_cai["data"][0] centroids_ptr = <uintptr_t>centroids_cai["data"][0] handle = handle if handle is not None else Handle() cdef device_resources *h = <device_resources*><size_t>handle.getHandle() x_c_contiguous = is_c_contiguous(x_cai) centroids_c_contiguous = is_c_contiguous(centroids_cai) if not x_c_contiguous or not centroids_c_contiguous: raise ValueError("Inputs must all be c contiguous") if not do_dtypes_match(X, centroids): raise ValueError("Inputs must all have the same dtypes " "(float32 or float64)") cdef float f_cost = 0 cdef double d_cost = 0 if x_dt == np.float32: cpp_cluster_cost(deref(h), <float*> x_ptr, <int> m, <int> x_k, <int> n_clusters, <float*> centroids_ptr, <float*> &f_cost) return f_cost elif x_dt == np.float64: cpp_cluster_cost(deref(h), <double*> x_ptr, <int> m, <int> x_k, <int> n_clusters, <double*> centroids_ptr, <double*> &d_cost) return d_cost else: raise ValueError("dtype %s not supported" % x_dt) class InitMethod(IntEnum): """ Method for initializing kmeans """ KMeansPlusPlus = <int> kmeans_types.InitMethod.KMeansPlusPlus Random = <int> kmeans_types.InitMethod.Random Array = <int> kmeans_types.InitMethod.Array cdef class KMeansParams: """ Specifies hyper-parameters for the kmeans algorithm. Parameters ---------- n_clusters : int, optional The number of clusters to form as well as the number of centroids to generate max_iter : int, optional Maximum number of iterations of the k-means algorithm for a single run tol : float, optional Relative tolerance with regards to inertia to declare convergence verbosity : int, optional seed: int, optional Seed to the random number generator. metric : str, optional Metric names to use for distance computation, see :func:`pylibraft.distance.pairwise_distance` for valid values. init : InitMethod, optional n_init : int, optional Number of instance k-means algorithm will be run with different seeds. oversampling_factor : float, optional Oversampling factor for use in the k-means algorithm """ cdef kmeans_types.KMeansParams c_obj def __init__(self, n_clusters: Optional[int] = None, max_iter: Optional[int] = None, tol: Optional[float] = None, verbosity: Optional[int] = None, seed: Optional[int] = None, metric: Optional[str] = None, init: Optional[InitMethod] = None, n_init: Optional[int] = None, oversampling_factor: Optional[float] = None, batch_samples: Optional[int] = None, batch_centroids: Optional[int] = None, inertia_check: Optional[bool] = None): if n_clusters is not None: self.c_obj.n_clusters = n_clusters if max_iter is not None: self.c_obj.max_iter = max_iter if tol is not None: self.c_obj.tol = tol if verbosity is not None: self.c_obj.verbosity = verbosity if seed is not None: self.c_obj.rng_state.seed = seed if metric is not None: distance = DISTANCE_TYPES.get(metric) if distance is None: valid_metrics = list(DISTANCE_TYPES.keys()) raise ValueError(f"Unknown metric '{metric}'. Valid values " f"are: {valid_metrics}") self.c_obj.metric = distance if init is not None: self.c_obj.init = init if n_init is not None: self.c_obj.n_init = n_init if oversampling_factor is not None: self.c_obj.oversampling_factor = oversampling_factor if batch_samples is not None: self.c_obj.batch_samples = batch_samples if batch_centroids is not None: self.c_obj.batch_centroids = batch_centroids if inertia_check is not None: self.c_obj.inertia_check = inertia_check @property def n_clusters(self): return self.c_obj.n_clusters @property def max_iter(self): return self.c_obj.max_iter @property def tol(self): return self.c_obj.tol @property def verbosity(self): return self.c_obj.verbosity @property def seed(self): return self.c_obj.rng_state.seed @property def init(self): return InitMethod(self.c_obj.init) @property def oversampling_factor(self): return self.c_obj.oversampling_factor @property def batch_samples(self): return self.c_obj.batch_samples @property def batch_centroids(self): return self.c_obj.batch_centroids @property def inertia_check(self): return self.c_obj.inertia_check FitOutput = namedtuple("FitOutput", "centroids inertia n_iter") @auto_sync_handle @auto_convert_output def fit( KMeansParams params, X, centroids=None, sample_weights=None, handle=None ): """ Find clusters with the k-means algorithm Parameters ---------- params : KMeansParams Parameters to use to fit KMeans model X : Input CUDA array interface compliant matrix shape (m, k) centroids : Optional writable CUDA array interface compliant matrix shape (n_clusters, k) sample_weights : Optional input CUDA array interface compliant matrix shape (n_clusters, 1) default: None {handle_docstring} Returns ------- centroids : raft.device_ndarray The computed centroids for each cluster inertia : float Sum of squared distances of samples to their closest cluster center n_iter : int The number of iterations used to fit the model Examples -------- >>> import cupy as cp >>> from pylibraft.cluster.kmeans import fit, KMeansParams >>> n_samples = 5000 >>> n_features = 50 >>> n_clusters = 3 >>> X = cp.random.random_sample((n_samples, n_features), ... dtype=cp.float32) >>> params = KMeansParams(n_clusters=n_clusters) >>> centroids, inertia, n_iter = fit(params, X) """ cdef device_resources *h = <device_resources*><size_t>handle.getHandle() cdef float f_inertia = 0.0 cdef double d_inertia = 0.0 cdef int n_iter = 0 cdef optional[device_vector_view[const double, int]] d_sample_weights cdef optional[device_vector_view[const float, int]] f_sample_weights X_cai = cai_wrapper(X) dtype = X_cai.dtype if centroids is None: centroids_shape = (params.n_clusters, X_cai.shape[1]) centroids = device_ndarray.empty(centroids_shape, dtype=dtype) centroids_cai = cai_wrapper(centroids) # validate inputs have are all c-contiguous, and have a consistent dtype # and expected shape X_cai.validate_shape_dtype(2) centroids_cai.validate_shape_dtype(2, dtype) if sample_weights is not None: sample_weights_cai = cai_wrapper(sample_weights) sample_weights_cai.validate_shape_dtype(1, dtype) if dtype == np.float64: if sample_weights is not None: d_sample_weights = make_device_vector_view( <const double *><uintptr_t>sample_weights_cai.data, <int>sample_weights_cai.shape[0]) cpp_kmeans.fit( deref(h), params.c_obj, make_device_matrix_view[double, int, row_major]( <double *><uintptr_t>X_cai.data, <int>X_cai.shape[0], <int>X_cai.shape[1]), d_sample_weights, make_device_matrix_view[double, int, row_major]( <double *><uintptr_t>centroids_cai.data, <int>centroids_cai.shape[0], <int>centroids_cai.shape[1]), make_host_scalar_view[double, int](&d_inertia), make_host_scalar_view[int, int](&n_iter)) return FitOutput(centroids, d_inertia, n_iter) elif dtype == np.float32: if sample_weights is not None: f_sample_weights = make_device_vector_view( <const float *><uintptr_t>sample_weights_cai.data, <int>sample_weights_cai.shape[0]) cpp_kmeans.fit( deref(h), params.c_obj, make_device_matrix_view[float, int, row_major]( <float *><uintptr_t>X_cai.data, <int>X_cai.shape[0], <int>X_cai.shape[1]), f_sample_weights, make_device_matrix_view[float, int, row_major]( <float *><uintptr_t>centroids_cai.data, <int>centroids_cai.shape[0], <int>centroids_cai.shape[1]), make_host_scalar_view[float, int](&f_inertia), make_host_scalar_view[int, int](&n_iter)) return FitOutput(centroids, f_inertia, n_iter) else: raise ValueError(f"unhandled dtype {dtype}")
0
rapidsai_public_repos/raft/python/pylibraft/pylibraft/cluster
rapidsai_public_repos/raft/python/pylibraft/pylibraft/cluster/cpp/kmeans_types.pxd
# # Copyright (c) 2022, NVIDIA CORPORATION. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. from libcpp cimport bool from pylibraft.distance.distance_type cimport DistanceType from pylibraft.random.cpp.rng_state cimport RngState cdef extern from "raft/cluster/kmeans_types.hpp" \ namespace "raft::cluster::kmeans": ctypedef enum InitMethod 'raft::cluster::KMeansParams::InitMethod': KMeansPlusPlus 'raft::cluster::kmeans::KMeansParams::InitMethod::KMeansPlusPlus' # noqa Random 'raft::cluster::kmeans::KMeansParams::InitMethod::Random' Array 'raft::cluster::kmeans::KMeansParams::InitMethod::Array' cdef cppclass KMeansParams: KMeansParams() except + int n_clusters InitMethod init int max_iter double tol int verbosity RngState rng_state DistanceType metric int n_init double oversampling_factor int batch_samples int batch_centroids bool inertia_check
0
rapidsai_public_repos/raft/python/pylibraft/pylibraft/cluster
rapidsai_public_repos/raft/python/pylibraft/pylibraft/cluster/cpp/kmeans.pxd
# # Copyright (c) 2022-2023, NVIDIA CORPORATION. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # # cython: profile=False # distutils: language = c++ # cython: embedsignature = True # cython: language_level = 3 import numpy as np from cython.operator cimport dereference as deref from libc.stdint cimport uintptr_t from libcpp cimport bool, nullptr from pylibraft.cluster.cpp.kmeans_types cimport KMeansParams from pylibraft.common.cpp.mdspan cimport * from pylibraft.common.cpp.optional cimport optional from pylibraft.common.handle cimport device_resources cdef extern from "raft_runtime/cluster/kmeans.hpp" \ namespace "raft::runtime::cluster::kmeans" nogil: cdef void update_centroids( const device_resources& handle, const double *X, int n_samples, int n_features, int n_clusters, const double *sample_weights, const double *centroids, const int* labels, double *new_centroids, double *weight_per_cluster) except + cdef void update_centroids( const device_resources& handle, const float *X, int n_samples, int n_features, int n_clusters, const float *sample_weights, const float *centroids, const int* labels, float *new_centroids, float *weight_per_cluster) except + cdef void cluster_cost( const device_resources& handle, const float* X, int n_samples, int n_features, int n_clusters, const float * centroids, float * cost) except + cdef void cluster_cost( const device_resources& handle, const double* X, int n_samples, int n_features, int n_clusters, const double * centroids, double * cost) except + cdef void init_plus_plus( const device_resources & handle, const KMeansParams& params, device_matrix_view[float, int, row_major] X, device_matrix_view[float, int, row_major] centroids) except + cdef void init_plus_plus( const device_resources & handle, const KMeansParams& params, device_matrix_view[double, int, row_major] X, device_matrix_view[double, int, row_major] centroids) except + cdef void fit( const device_resources & handle, const KMeansParams& params, device_matrix_view[float, int, row_major] X, optional[device_vector_view[float, int]] sample_weight, device_matrix_view[float, int, row_major] inertia, host_scalar_view[float, int] inertia, host_scalar_view[int, int] n_iter) except + cdef void fit( const device_resources & handle, const KMeansParams& params, device_matrix_view[double, int, row_major] X, optional[device_vector_view[double, int]] sample_weight, device_matrix_view[double, int, row_major] inertia, host_scalar_view[double, int] inertia, host_scalar_view[int, int] n_iter) except +
0
rapidsai_public_repos/raft/conda
rapidsai_public_repos/raft/conda/environments/bench_ann_cuda-118_arch-x86_64.yaml
# This file is generated by `rapids-dependency-file-generator`. # To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`. channels: - rapidsai - rapidsai-nightly - dask/label/dev - conda-forge - nvidia dependencies: - benchmark>=1.8.2 - c-compiler - clang-tools=16.0.6 - clang==16.0.6 - cmake>=3.26.4 - cuda-nvtx=11.8 - cuda-profiler-api=11.8.86 - cuda-version=11.8 - cudatoolkit - cxx-compiler - cython>=3.0.0 - gcc_linux-64=11.* - glog>=0.6.0 - h5py>=3.8.0 - hnswlib=0.7.0 - libcublas-dev=11.11.3.6 - libcublas=11.11.3.6 - libcurand-dev=10.3.0.86 - libcurand=10.3.0.86 - libcusolver-dev=11.4.1.48 - libcusolver=11.4.1.48 - libcusparse-dev=11.7.5.86 - libcusparse=11.7.5.86 - matplotlib - nccl>=2.9.9 - ninja - nlohmann_json>=3.11.2 - nvcc_linux-64=11.8 - openblas - pandas - pyyaml - rmm==24.2.* - scikit-build>=0.13.1 - sysroot_linux-64==2.17 name: bench_ann_cuda-118_arch-x86_64
0
rapidsai_public_repos/raft/conda
rapidsai_public_repos/raft/conda/environments/bench_ann_cuda-118_arch-aarch64.yaml
# This file is generated by `rapids-dependency-file-generator`. # To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`. channels: - rapidsai - rapidsai-nightly - dask/label/dev - conda-forge - nvidia dependencies: - benchmark>=1.8.2 - c-compiler - clang-tools=16.0.6 - clang==16.0.6 - cmake>=3.26.4 - cuda-nvtx=11.8 - cuda-profiler-api=11.8.86 - cuda-version=11.8 - cudatoolkit - cxx-compiler - cython>=3.0.0 - gcc_linux-aarch64=11.* - glog>=0.6.0 - h5py>=3.8.0 - hnswlib=0.7.0 - libcublas-dev=11.11.3.6 - libcublas=11.11.3.6 - libcurand-dev=10.3.0.86 - libcurand=10.3.0.86 - libcusolver-dev=11.4.1.48 - libcusolver=11.4.1.48 - libcusparse-dev=11.7.5.86 - libcusparse=11.7.5.86 - matplotlib - nccl>=2.9.9 - ninja - nlohmann_json>=3.11.2 - nvcc_linux-aarch64=11.8 - openblas - pandas - pyyaml - rmm==24.2.* - scikit-build>=0.13.1 - sysroot_linux-aarch64==2.17 name: bench_ann_cuda-118_arch-aarch64
0
rapidsai_public_repos/raft/conda
rapidsai_public_repos/raft/conda/environments/bench_ann_cuda-120_arch-aarch64.yaml
# This file is generated by `rapids-dependency-file-generator`. # To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`. channels: - rapidsai - rapidsai-nightly - dask/label/dev - conda-forge - nvidia dependencies: - benchmark>=1.8.2 - c-compiler - clang-tools=16.0.6 - clang==16.0.6 - cmake>=3.26.4 - cuda-cudart-dev - cuda-nvcc - cuda-nvtx-dev - cuda-profiler-api - cuda-version=12.0 - cxx-compiler - cython>=3.0.0 - gcc_linux-aarch64=11.* - glog>=0.6.0 - h5py>=3.8.0 - hnswlib=0.7.0 - libcublas-dev - libcurand-dev - libcusolver-dev - libcusparse-dev - matplotlib - nccl>=2.9.9 - ninja - nlohmann_json>=3.11.2 - openblas - pandas - pyyaml - rmm==24.2.* - scikit-build>=0.13.1 - sysroot_linux-aarch64==2.17 name: bench_ann_cuda-120_arch-aarch64
0
rapidsai_public_repos/raft/conda
rapidsai_public_repos/raft/conda/environments/all_cuda-120_arch-x86_64.yaml
# This file is generated by `rapids-dependency-file-generator`. # To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`. channels: - rapidsai - rapidsai-nightly - dask/label/dev - conda-forge - nvidia dependencies: - breathe - c-compiler - clang-tools=16.0.6 - clang==16.0.6 - cmake>=3.26.4 - cuda-cudart-dev - cuda-nvcc - cuda-nvtx-dev - cuda-profiler-api - cuda-python>=12.0,<13.0a0 - cuda-version=12.0 - cupy>=12.0.0 - cxx-compiler - cython>=3.0.0 - dask-cuda==24.2.* - doxygen>=1.8.20 - gcc_linux-64=11.* - gmock>=1.13.0 - graphviz - gtest>=1.13.0 - ipython - joblib>=0.11 - libcublas-dev - libcurand-dev - libcusolver-dev - libcusparse-dev - nccl>=2.9.9 - ninja - numba>=0.57 - numpy>=1.21 - numpydoc - pre-commit - pydata-sphinx-theme - pytest - pytest-cov - rapids-dask-dependency==24.2.* - recommonmark - rmm==24.2.* - scikit-build>=0.13.1 - scikit-learn - scipy - sphinx-copybutton - sphinx-markdown-tables - sysroot_linux-64==2.17 - ucx-proc=*=gpu - ucx-py==0.36.* - ucx>=1.13.0 name: all_cuda-120_arch-x86_64
0
rapidsai_public_repos/raft/conda
rapidsai_public_repos/raft/conda/environments/all_cuda-120_arch-aarch64.yaml
# This file is generated by `rapids-dependency-file-generator`. # To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`. channels: - rapidsai - rapidsai-nightly - dask/label/dev - conda-forge - nvidia dependencies: - breathe - c-compiler - clang-tools=16.0.6 - clang==16.0.6 - cmake>=3.26.4 - cuda-cudart-dev - cuda-nvcc - cuda-nvtx-dev - cuda-profiler-api - cuda-python>=12.0,<13.0a0 - cuda-version=12.0 - cupy>=12.0.0 - cxx-compiler - cython>=3.0.0 - dask-cuda==24.2.* - doxygen>=1.8.20 - gcc_linux-aarch64=11.* - gmock>=1.13.0 - graphviz - gtest>=1.13.0 - ipython - joblib>=0.11 - libcublas-dev - libcurand-dev - libcusolver-dev - libcusparse-dev - nccl>=2.9.9 - ninja - numba>=0.57 - numpy>=1.21 - numpydoc - pre-commit - pydata-sphinx-theme - pytest - pytest-cov - rapids-dask-dependency==24.2.* - recommonmark - rmm==24.2.* - scikit-build>=0.13.1 - scikit-learn - scipy - sphinx-copybutton - sphinx-markdown-tables - sysroot_linux-aarch64==2.17 - ucx-proc=*=gpu - ucx-py==0.36.* - ucx>=1.13.0 name: all_cuda-120_arch-aarch64
0
rapidsai_public_repos/raft/conda
rapidsai_public_repos/raft/conda/environments/all_cuda-118_arch-aarch64.yaml
# This file is generated by `rapids-dependency-file-generator`. # To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`. channels: - rapidsai - rapidsai-nightly - dask/label/dev - conda-forge - nvidia dependencies: - breathe - c-compiler - clang-tools=16.0.6 - clang==16.0.6 - cmake>=3.26.4 - cuda-nvtx=11.8 - cuda-profiler-api=11.8.86 - cuda-python>=11.7.1,<12.0a0 - cuda-version=11.8 - cudatoolkit - cupy>=12.0.0 - cxx-compiler - cython>=3.0.0 - dask-cuda==24.2.* - doxygen>=1.8.20 - gcc_linux-aarch64=11.* - gmock>=1.13.0 - graphviz - gtest>=1.13.0 - ipython - joblib>=0.11 - libcublas-dev=11.11.3.6 - libcublas=11.11.3.6 - libcurand-dev=10.3.0.86 - libcurand=10.3.0.86 - libcusolver-dev=11.4.1.48 - libcusolver=11.4.1.48 - libcusparse-dev=11.7.5.86 - libcusparse=11.7.5.86 - nccl>=2.9.9 - ninja - numba>=0.57 - numpy>=1.21 - numpydoc - nvcc_linux-aarch64=11.8 - pre-commit - pydata-sphinx-theme - pytest - pytest-cov - rapids-dask-dependency==24.2.* - recommonmark - rmm==24.2.* - scikit-build>=0.13.1 - scikit-learn - scipy - sphinx-copybutton - sphinx-markdown-tables - sysroot_linux-aarch64==2.17 - ucx-proc=*=gpu - ucx-py==0.36.* - ucx>=1.13.0 name: all_cuda-118_arch-aarch64
0
rapidsai_public_repos/raft/conda
rapidsai_public_repos/raft/conda/environments/all_cuda-118_arch-x86_64.yaml
# This file is generated by `rapids-dependency-file-generator`. # To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`. channels: - rapidsai - rapidsai-nightly - dask/label/dev - conda-forge - nvidia dependencies: - breathe - c-compiler - clang-tools=16.0.6 - clang==16.0.6 - cmake>=3.26.4 - cuda-nvtx=11.8 - cuda-profiler-api=11.8.86 - cuda-python>=11.7.1,<12.0a0 - cuda-version=11.8 - cudatoolkit - cupy>=12.0.0 - cxx-compiler - cython>=3.0.0 - dask-cuda==24.2.* - doxygen>=1.8.20 - gcc_linux-64=11.* - gmock>=1.13.0 - graphviz - gtest>=1.13.0 - ipython - joblib>=0.11 - libcublas-dev=11.11.3.6 - libcublas=11.11.3.6 - libcurand-dev=10.3.0.86 - libcurand=10.3.0.86 - libcusolver-dev=11.4.1.48 - libcusolver=11.4.1.48 - libcusparse-dev=11.7.5.86 - libcusparse=11.7.5.86 - nccl>=2.9.9 - ninja - numba>=0.57 - numpy>=1.21 - numpydoc - nvcc_linux-64=11.8 - pre-commit - pydata-sphinx-theme - pytest - pytest-cov - rapids-dask-dependency==24.2.* - recommonmark - rmm==24.2.* - scikit-build>=0.13.1 - scikit-learn - scipy - sphinx-copybutton - sphinx-markdown-tables - sysroot_linux-64==2.17 - ucx-proc=*=gpu - ucx-py==0.36.* - ucx>=1.13.0 name: all_cuda-118_arch-x86_64
0
rapidsai_public_repos/raft/conda
rapidsai_public_repos/raft/conda/environments/bench_ann_cuda-120_arch-x86_64.yaml
# This file is generated by `rapids-dependency-file-generator`. # To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`. channels: - rapidsai - rapidsai-nightly - dask/label/dev - conda-forge - nvidia dependencies: - benchmark>=1.8.2 - c-compiler - clang-tools=16.0.6 - clang==16.0.6 - cmake>=3.26.4 - cuda-cudart-dev - cuda-nvcc - cuda-nvtx-dev - cuda-profiler-api - cuda-version=12.0 - cxx-compiler - cython>=3.0.0 - gcc_linux-64=11.* - glog>=0.6.0 - h5py>=3.8.0 - hnswlib=0.7.0 - libcublas-dev - libcurand-dev - libcusolver-dev - libcusparse-dev - matplotlib - nccl>=2.9.9 - ninja - nlohmann_json>=3.11.2 - openblas - pandas - pyyaml - rmm==24.2.* - scikit-build>=0.13.1 - sysroot_linux-64==2.17 name: bench_ann_cuda-120_arch-x86_64
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/raft-ann-bench-cpu/conda_build_config.yaml
c_compiler_version: - 11 cxx_compiler_version: - 11 sysroot_version: - "2.17" cmake_version: - ">=3.26.4" glog_version: - ">=0.6.0" h5py_version: - ">=3.8.0" nlohmann_json_version: - ">=3.11.2" spdlog_version: - ">=1.11.0,<1.12" fmt_version: - ">=9.1.0,<10"
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/raft-ann-bench-cpu/build.sh
#!/usr/bin/env bash # Copyright (c) 2023, NVIDIA CORPORATION. ./build.sh bench-ann --cpu-only --no-nvtx --build-metrics=bench_ann_cpu --incl-cache-stats cmake --install cpp/build --component ann_bench
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/raft-ann-bench-cpu/meta.yaml
# Copyright (c) 2022-2023, NVIDIA CORPORATION. # Usage: # conda build . -c conda-forge -c nvidia -c rapidsai {% set version = environ['RAPIDS_PACKAGE_VERSION'].lstrip('v') + environ.get('VERSION_SUFFIX', '') %} {% set minor_version = version.split('.')[0] + '.' + version.split('.')[1] %} {% set py_version = environ['CONDA_PY'] %} {% set cuda_version = '.'.join(environ['RAPIDS_CUDA_VERSION'].split('.')[:2]) %} {% set date_string = environ['RAPIDS_DATE_STRING'] %} package: name: raft-ann-bench-cpu version: {{ version }} script: build.sh source: path: ../../.. build: script_env: - AWS_ACCESS_KEY_ID - AWS_SECRET_ACCESS_KEY - AWS_SESSION_TOKEN - CMAKE_C_COMPILER_LAUNCHER - CMAKE_CUDA_COMPILER_LAUNCHER - CMAKE_CXX_COMPILER_LAUNCHER - CMAKE_GENERATOR - PARALLEL_LEVEL - RAPIDS_ARTIFACTS_DIR - SCCACHE_BUCKET - SCCACHE_IDLE_TIMEOUT - SCCACHE_REGION - SCCACHE_S3_KEY_PREFIX=libraft-aarch64 # [aarch64] - SCCACHE_S3_KEY_PREFIX=libraft-linux64 # [linux64] - SCCACHE_S3_USE_SSL number: {{ GIT_DESCRIBE_NUMBER }} string: py{{ py_version }}_{{ date_string }}_{{ GIT_DESCRIBE_HASH }}_{{ GIT_DESCRIBE_NUMBER }} requirements: build: - {{ compiler('c') }} - {{ compiler('cxx') }} - cmake {{ cmake_version }} - ninja - sysroot_{{ target_platform }} {{ sysroot_version }} host: - glog {{ glog_version }} - matplotlib - nlohmann_json {{ nlohmann_json_version }} - spdlog {{ spdlog_version }} - fmt {{ fmt_version }} - python - pyyaml - pandas run: - glog {{ glog_version }} - h5py {{ h5py_version }} - matplotlib - python - pyyaml - pandas - benchmark about: home: https://rapids.ai/ license: Apache-2.0 summary: RAFT ANN CPU benchmarks
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/raft-dask/conda_build_config.yaml
c_compiler_version: - 11 cxx_compiler_version: - 11 cuda_compiler: - cuda-nvcc cuda11_compiler: - nvcc sysroot_version: - "2.17" ucx_version: - ">=1.14.1,<1.16.0" ucx_py_version: - "0.36.*" cmake_version: - ">=3.26.4"
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/raft-dask/build.sh
# Copyright (c) 2022, NVIDIA CORPORATION. #!/usr/bin/env bash # This assumes the script is executed from the root of the repo directory ./build.sh raft-dask --no-nvtx
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/raft-dask/meta.yaml
# Copyright (c) 2022-2023, NVIDIA CORPORATION. # Usage: # conda build . -c conda-forge -c numba -c rapidsai -c pytorch {% set version = environ['RAPIDS_PACKAGE_VERSION'].lstrip('v') + environ.get('VERSION_SUFFIX', '') %} {% set minor_version = version.split('.')[0] + '.' + version.split('.')[1] %} {% set py_version = environ['CONDA_PY'] %} {% set cuda_version = '.'.join(environ['RAPIDS_CUDA_VERSION'].split('.')[:2]) %} {% set cuda_major = cuda_version.split('.')[0] %} {% set date_string = environ['RAPIDS_DATE_STRING'] %} package: name: raft-dask version: {{ version }} source: path: ../../.. build: number: {{ GIT_DESCRIBE_NUMBER }} string: cuda{{ cuda_major }}_py{{ py_version }}_{{ date_string }}_{{ GIT_DESCRIBE_HASH }}_{{ GIT_DESCRIBE_NUMBER }} ignore_run_exports_from: {% if cuda_major == "11" %} - {{ compiler('cuda11') }} {% endif %} requirements: build: - {{ compiler('c') }} - {{ compiler('cxx') }} {% if cuda_major == "11" %} - {{ compiler('cuda11') }} ={{ cuda_version }} {% else %} - {{ compiler('cuda') }} {% endif %} - cuda-version ={{ cuda_version }} - cmake {{ cmake_version }} - ninja - sysroot_{{ target_platform }} {{ sysroot_version }} host: {% if cuda_major == "11" %} - cuda-python >=11.7.1,<12.0a0 - cudatoolkit {% else %} - cuda-python >=12.0,<13.0a0 {% endif %} - cuda-version ={{ cuda_version }} - cython >=3.0.0 - nccl >=2.9.9 - pylibraft {{ version }} - python x.x - rmm ={{ minor_version }} - scikit-build >=0.13.1 - setuptools - ucx {{ ucx_version }} - ucx-proc=*=gpu - ucx-py {{ ucx_py_version }} run: {% if cuda_major == "11" %} - cudatoolkit {% endif %} - {{ pin_compatible('cuda-version', max_pin='x', min_pin='x') }} - dask-cuda ={{ minor_version }} - rapids-dask-dependency ={{ minor_version }} - joblib >=0.11 - nccl >=2.9.9 - pylibraft {{ version }} - python x.x - rmm ={{ minor_version }} - ucx {{ ucx_version }} - ucx-proc=*=gpu - ucx-py {{ ucx_py_version }} tests: requirements: - cuda-version ={{ cuda_version }} imports: - raft_dask about: home: https://rapids.ai/ license: Apache-2.0 # license_file: LICENSE summary: raft-dask library
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/libraft/conda_build_config.yaml
c_compiler_version: - 11 cxx_compiler_version: - 11 cuda_compiler: - cuda-nvcc cuda11_compiler: - nvcc sysroot_version: - "2.17" cmake_version: - ">=3.26.4" nccl_version: - ">=2.9.9" gtest_version: - ">=1.13.0" glog_version: - ">=0.6.0" faiss_version: - ">=1.7.1" h5py_version: - ">=3.8.0" nlohmann_json_version: - ">=3.11.2" # The CTK libraries below are missing from the conda-forge::cudatoolkit package # for CUDA 11. The "*_host_*" version specifiers correspond to `11.8` packages # and the "*_run_*" version specifiers correspond to `11.x` packages. cuda11_libcublas_host_version: - "=11.11.3.6" cuda11_libcublas_run_version: - ">=11.5.2.43,<12.0.0" cuda11_libcurand_host_version: - "=10.3.0.86" cuda11_libcurand_run_version: - ">=10.2.5.43,<10.3.1" cuda11_libcusolver_host_version: - "=11.4.1.48" cuda11_libcusolver_run_version: - ">=11.2.0.43,<11.4.2" cuda11_libcusparse_host_version: - "=11.7.5.86" cuda11_libcusparse_run_version: - ">=11.6.0.43,<12.0.0" # `cuda-profiler-api` only has `11.8.0` and `12.0.0` packages for all # architectures. The "*_host_*" version specifiers correspond to `11.8` packages and the # "*_run_*" version specifiers correspond to `11.x` packages. cuda11_cuda_profiler_api_host_version: - "=11.8.86" cuda11_cuda_profiler_api_run_version: - ">=11.4.240,<12" spdlog_version: - ">=1.11.0,<1.12" fmt_version: - ">=9.1.0,<10"
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/libraft/build_libraft_template.sh
#!/usr/bin/env bash # Copyright (c) 2022-2023, NVIDIA CORPORATION. # Just building template so we verify it uses libraft.so and fail if it doesn't build ./build.sh template
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/libraft/build_libraft_tests.sh
#!/usr/bin/env bash # Copyright (c) 2022-2023, NVIDIA CORPORATION. ./build.sh tests bench-prims --allgpuarch --no-nvtx --build-metrics=tests_bench_prims --incl-cache-stats cmake --install cpp/build --component testing
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/libraft/meta.yaml
# Copyright (c) 2022-2023, NVIDIA CORPORATION. # Usage: # conda build . -c conda-forge -c nvidia -c rapidsai {% set version = environ['RAPIDS_PACKAGE_VERSION'].lstrip('v') + environ.get('VERSION_SUFFIX', '') %} {% set minor_version = version.split('.')[0] + '.' + version.split('.')[1] %} {% set cuda_version = '.'.join(environ['RAPIDS_CUDA_VERSION'].split('.')[:2]) %} {% set cuda_major = cuda_version.split('.')[0] %} {% set cuda_spec = ">=" + cuda_major ~ ",<" + (cuda_major | int + 1) ~ ".0a0" %} # i.e. >=11,<12.0a0 {% set date_string = environ['RAPIDS_DATE_STRING'] %} package: name: libraft-split source: path: ../../.. outputs: - name: libraft-headers-only version: {{ version }} script: build_libraft_headers.sh build: script_env: &script_env - AWS_ACCESS_KEY_ID - AWS_SECRET_ACCESS_KEY - AWS_SESSION_TOKEN - CMAKE_C_COMPILER_LAUNCHER - CMAKE_CUDA_COMPILER_LAUNCHER - CMAKE_CXX_COMPILER_LAUNCHER - CMAKE_GENERATOR - PARALLEL_LEVEL - RAPIDS_ARTIFACTS_DIR - SCCACHE_BUCKET - SCCACHE_IDLE_TIMEOUT - SCCACHE_REGION - SCCACHE_S3_KEY_PREFIX=libraft-aarch64 # [aarch64] - SCCACHE_S3_KEY_PREFIX=libraft-linux64 # [linux64] - SCCACHE_S3_USE_SSL - SCCACHE_S3_NO_CREDENTIALS number: {{ GIT_DESCRIBE_NUMBER }} string: cuda{{ cuda_major }}_{{ date_string }}_{{ GIT_DESCRIBE_HASH }}_{{ GIT_DESCRIBE_NUMBER }} ignore_run_exports_from: {% if cuda_major == "11" %} - {{ compiler('cuda11') }} {% endif %} - librmm requirements: build: - {{ compiler('c') }} - {{ compiler('cxx') }} {% if cuda_major == "11" %} - {{ compiler('cuda11') }} ={{ cuda_version }} {% else %} - {{ compiler('cuda') }} {% endif %} - cuda-version ={{ cuda_version }} - cmake {{ cmake_version }} - ninja - sysroot_{{ target_platform }} {{ sysroot_version }} host: {% if cuda_major != "11" %} - cuda-cudart-dev {% endif %} - cuda-version ={{ cuda_version }} - librmm ={{ minor_version }} - spdlog {{ spdlog_version }} - fmt {{ fmt_version }} run: - {{ pin_compatible('cuda-version', max_pin='x', min_pin='x') }} {% if cuda_major == "11" %} - cudatoolkit {% endif %} - librmm ={{ minor_version }} - spdlog {{ spdlog_version }} - fmt {{ fmt_version }} about: home: https://rapids.ai/ license: Apache-2.0 summary: libraft-headers-only library - name: libraft-headers version: {{ version }} build: number: {{ GIT_DESCRIBE_NUMBER }} string: cuda{{ cuda_major }}_{{ date_string }}_{{ GIT_DESCRIBE_HASH }}_{{ GIT_DESCRIBE_NUMBER }} ignore_run_exports_from: {% if cuda_major == "11" %} - {{ compiler('cuda11') }} {% endif %} - librmm requirements: host: - cuda-version ={{ cuda_version }} run: - {{ pin_subpackage('libraft-headers-only', exact=True) }} - librmm ={{ minor_version }} - {{ pin_compatible('cuda-version', max_pin='x', min_pin='x') }} {% if cuda_major == "11" %} - cudatoolkit - cuda-profiler-api {{ cuda11_cuda_profiler_api_run_version }} - libcublas {{ cuda11_libcublas_run_version }} - libcublas-dev {{ cuda11_libcublas_run_version }} - libcurand {{ cuda11_libcurand_run_version }} - libcurand-dev {{ cuda11_libcurand_run_version }} - libcusolver {{ cuda11_libcusolver_run_version }} - libcusolver-dev {{ cuda11_libcusolver_run_version }} - libcusparse {{ cuda11_libcusparse_run_version }} - libcusparse-dev {{ cuda11_libcusparse_run_version }} {% else %} - cuda-cudart-dev - cuda-profiler-api - libcublas-dev - libcurand-dev - libcusolver-dev - libcusparse-dev {% endif %} about: home: https://rapids.ai/ license: Apache-2.0 summary: libraft-headers library - name: libraft version: {{ version }} script: build_libraft.sh build: script_env: *script_env number: {{ GIT_DESCRIBE_NUMBER }} string: cuda{{ cuda_major }}_{{ date_string }}_{{ GIT_DESCRIBE_HASH }}_{{ GIT_DESCRIBE_NUMBER }} ignore_run_exports_from: {% if cuda_major == "11" %} - {{ compiler('cuda11') }} {% endif %} requirements: build: - {{ compiler('c') }} - {{ compiler('cxx') }} {% if cuda_major == "11" %} - {{ compiler('cuda11') }} ={{ cuda_version }} {% else %} - {{ compiler('cuda') }} {% endif %} - cuda-version ={{ cuda_version }} - cmake {{ cmake_version }} - ninja - sysroot_{{ target_platform }} {{ sysroot_version }} host: - {{ pin_subpackage('libraft-headers', exact=True) }} - cuda-version ={{ cuda_version }} {% if cuda_major == "11" %} - cuda-profiler-api {{ cuda11_cuda_profiler_api_host_version }} - libcublas {{ cuda11_libcublas_host_version }} - libcublas-dev {{ cuda11_libcublas_host_version }} - libcurand {{ cuda11_libcurand_host_version }} - libcurand-dev {{ cuda11_libcurand_host_version }} - libcusolver {{ cuda11_libcusolver_host_version }} - libcusolver-dev {{ cuda11_libcusolver_host_version }} - libcusparse {{ cuda11_libcusparse_host_version }} - libcusparse-dev {{ cuda11_libcusparse_host_version }} {% else %} - cuda-profiler-api - libcublas-dev - libcurand-dev - libcusolver-dev - libcusparse-dev {% endif %} run: - {{ pin_subpackage('libraft-headers', exact=True) }} - {{ pin_compatible('cuda-version', max_pin='x', min_pin='x') }} about: home: https://rapids.ai/ license: Apache-2.0 summary: libraft library - name: libraft-static version: {{ version }} script: build_libraft_static.sh build: script_env: *script_env number: {{ GIT_DESCRIBE_NUMBER }} string: cuda{{ cuda_major }}_{{ date_string }}_{{ GIT_DESCRIBE_HASH }}_{{ GIT_DESCRIBE_NUMBER }} ignore_run_exports_from: {% if cuda_major == "11" %} - {{ compiler('cuda11') }} {% endif %} requirements: build: - {{ compiler('c') }} - {{ compiler('cxx') }} {% if cuda_major == "11" %} - {{ compiler('cuda11') }} ={{ cuda_version }} {% else %} - {{ compiler('cuda') }} {% endif %} - cuda-version ={{ cuda_version }} - cmake {{ cmake_version }} - ninja - sysroot_{{ target_platform }} {{ sysroot_version }} host: - {{ pin_subpackage('libraft-headers', exact=True) }} - cuda-version ={{ cuda_version }} {% if cuda_major == "11" %} - cuda-profiler-api {{ cuda11_cuda_profiler_api_host_version }} - libcublas {{ cuda11_libcublas_host_version }} - libcublas-dev {{ cuda11_libcublas_host_version }} - libcurand {{ cuda11_libcurand_host_version }} - libcurand-dev {{ cuda11_libcurand_host_version }} - libcusolver {{ cuda11_libcusolver_host_version }} - libcusolver-dev {{ cuda11_libcusolver_host_version }} - libcusparse {{ cuda11_libcusparse_host_version }} - libcusparse-dev {{ cuda11_libcusparse_host_version }} {% else %} - cuda-profiler-api - libcublas-dev - libcurand-dev - libcusolver-dev - libcusparse-dev {% endif %} run: - {{ pin_subpackage('libraft-headers', exact=True) }} - {{ pin_compatible('cuda-version', max_pin='x', min_pin='x') }} about: home: https://rapids.ai/ license: Apache-2.0 summary: libraft static library - name: libraft-tests version: {{ version }} script: build_libraft_tests.sh build: script_env: *script_env number: {{ GIT_DESCRIBE_NUMBER }} string: cuda{{ cuda_major }}_{{ date_string }}_{{ GIT_DESCRIBE_HASH }}_{{ GIT_DESCRIBE_NUMBER }} ignore_run_exports_from: {% if cuda_major == "11" %} - {{ compiler('cuda11') }} {% endif %} requirements: build: - {{ compiler('c') }} - {{ compiler('cxx') }} {% if cuda_major == "11" %} - {{ compiler('cuda11') }} ={{ cuda_version }} {% else %} - {{ compiler('cuda') }} {% endif %} - cuda-version ={{ cuda_version }} - cmake {{ cmake_version }} - ninja - sysroot_{{ target_platform }} {{ sysroot_version }} host: - {{ pin_subpackage('libraft', exact=True) }} - cuda-version ={{ cuda_version }} {% if cuda_major == "11" %} - cuda-profiler-api {{ cuda11_cuda_profiler_api_run_version }} - libcublas {{ cuda11_libcublas_host_version }} - libcublas-dev {{ cuda11_libcublas_host_version }} - libcurand {{ cuda11_libcurand_host_version }} - libcurand-dev {{ cuda11_libcurand_host_version }} - libcusolver {{ cuda11_libcusolver_host_version }} - libcusolver-dev {{ cuda11_libcusolver_host_version }} - libcusparse {{ cuda11_libcusparse_host_version }} - libcusparse-dev {{ cuda11_libcusparse_host_version }} {% else %} - cuda-cudart-dev - cuda-profiler-api - libcublas-dev - libcurand-dev - libcusolver-dev - libcusparse-dev {% endif %} - gmock {{ gtest_version }} - gtest {{ gtest_version }} run: - {{ pin_compatible('cuda-version', max_pin='x', min_pin='x') }} {% if cuda_major == "11" %} - cudatoolkit {% endif %} - {{ pin_subpackage('libraft', exact=True) }} - gmock {{ gtest_version }} - gtest {{ gtest_version }} about: home: https://rapids.ai/ license: Apache-2.0 summary: libraft tests - name: libraft-template version: {{ version }} script: build_libraft_template.sh build: script_env: *script_env number: {{ GIT_DESCRIBE_NUMBER }} string: cuda{{ cuda_major }}_{{ date_string }}_{{ GIT_DESCRIBE_HASH }}_{{ GIT_DESCRIBE_NUMBER }} ignore_run_exports_from: {% if cuda_major == "11" %} - {{ compiler('cuda11') }} {% endif %} requirements: build: - {{ compiler('c') }} - {{ compiler('cxx') }} {% if cuda_major == "11" %} - {{ compiler('cuda11') }} ={{ cuda_version }} {% else %} - {{ compiler('cuda') }} {% endif %} - cuda-version ={{ cuda_version }} - cmake {{ cmake_version }} - ninja - sysroot_{{ target_platform }} {{ sysroot_version }} host: - {{ pin_subpackage('libraft', exact=True) }} - {{ pin_subpackage('libraft-headers', exact=True) }} - cuda-version ={{ cuda_version }} {% if cuda_major == "11" %} - cuda-profiler-api {{ cuda11_cuda_profiler_api_run_version }} - libcublas {{ cuda11_libcublas_host_version }} - libcublas-dev {{ cuda11_libcublas_host_version }} {% else %} - cuda-profiler-api - libcublas-dev {% endif %} run: - {{ pin_compatible('cuda-version', max_pin='x', min_pin='x') }} {% if cuda_major == "11" %} - cudatoolkit {% endif %} - {{ pin_subpackage('libraft', exact=True) }} about: home: https://rapids.ai/ license: Apache-2.0 summary: libraft template
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/libraft/build_libraft.sh
#!/usr/bin/env bash # Copyright (c) 2022-2023, NVIDIA CORPORATION. ./build.sh libraft --allgpuarch --compile-lib --build-metrics=compile_lib --incl-cache-stats --no-nvtx
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/libraft/build_libraft_headers.sh
#!/usr/bin/env bash # Copyright (c) 2022-2023, NVIDIA CORPORATION. ./build.sh libraft --allgpuarch --no-nvtx
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/libraft/build_libraft_static.sh
#!/usr/bin/env bash # Copyright (c) 2022-2023, NVIDIA CORPORATION. ./build.sh libraft --allgpuarch --compile-static-lib --build-metrics=compile_lib_static --incl-cache-stats --no-nvtx -n cmake --install cpp/build --component compiled-static
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/raft-ann-bench/conda_build_config.yaml
c_compiler_version: - 11 cxx_compiler_version: - 11 cuda_compiler: - cuda-nvcc cuda11_compiler: - nvcc sysroot_version: - "2.17" cmake_version: - ">=3.26.4" nccl_version: - ">=2.9.9" gtest_version: - ">=1.13.0" glog_version: - ">=0.6.0" h5py_version: - ">=3.8.0" nlohmann_json_version: - ">=3.11.2" # The CTK libraries below are missing from the conda-forge::cudatoolkit package # for CUDA 11. The "*_host_*" version specifiers correspond to `11.8` packages # and the "*_run_*" version specifiers correspond to `11.x` packages. cuda11_libcublas_host_version: - "=11.11.3.6" cuda11_libcublas_run_version: - ">=11.5.2.43,<12.0.0" cuda11_libcurand_host_version: - "=10.3.0.86" cuda11_libcurand_run_version: - ">=10.2.5.43,<10.3.1" cuda11_libcusolver_host_version: - "=11.4.1.48" cuda11_libcusolver_run_version: - ">=11.2.0.43,<11.4.2" cuda11_libcusparse_host_version: - "=11.7.5.86" cuda11_libcusparse_run_version: - ">=11.6.0.43,<12.0.0" # `cuda-profiler-api` only has `11.8.0` and `12.0.0` packages for all # architectures. The "*_host_*" version specifiers correspond to `11.8` packages and the # "*_run_*" version specifiers correspond to `11.x` packages. cuda11_cuda_profiler_api_host_version: - "=11.8.86" cuda11_cuda_profiler_api_run_version: - ">=11.4.240,<12"
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/raft-ann-bench/build.sh
#!/usr/bin/env bash # Copyright (c) 2023, NVIDIA CORPORATION. ./build.sh bench-ann --allgpuarch --no-nvtx --build-metrics=bench_ann --incl-cache-stats cmake --install cpp/build --component ann_bench
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/raft-ann-bench/meta.yaml
# Copyright (c) 2022-2023, NVIDIA CORPORATION. # Usage: # conda build . -c conda-forge -c nvidia -c rapidsai {% set version = environ['RAPIDS_PACKAGE_VERSION'].lstrip('v') + environ.get('VERSION_SUFFIX', '') %} {% set minor_version = version.split('.')[0] + '.' + version.split('.')[1] %} {% set py_version = environ['CONDA_PY'] %} {% set cuda_version = '.'.join(environ['RAPIDS_CUDA_VERSION'].split('.')[:2]) %} {% set cuda_major = cuda_version.split('.')[0] %} {% set cuda_spec = ">=" + cuda_major ~ ",<" + (cuda_major | int + 1) ~ ".0a0" %} # i.e. >=11,<12.0a0 {% set date_string = environ['RAPIDS_DATE_STRING'] %} package: name: raft-ann-bench version: {{ version }} script: build.sh source: path: ../../.. build: script_env: - AWS_ACCESS_KEY_ID - AWS_SECRET_ACCESS_KEY - AWS_SESSION_TOKEN - CMAKE_C_COMPILER_LAUNCHER - CMAKE_CUDA_COMPILER_LAUNCHER - CMAKE_CXX_COMPILER_LAUNCHER - CMAKE_GENERATOR - PARALLEL_LEVEL - RAPIDS_ARTIFACTS_DIR - SCCACHE_BUCKET - SCCACHE_IDLE_TIMEOUT - SCCACHE_REGION - SCCACHE_S3_KEY_PREFIX=libraft-aarch64 # [aarch64] - SCCACHE_S3_KEY_PREFIX=libraft-linux64 # [linux64] - SCCACHE_S3_USE_SSL number: {{ GIT_DESCRIBE_NUMBER }} string: cuda{{ cuda_major }}_py{{ py_version }}_{{ date_string }}_{{ GIT_DESCRIBE_HASH }}_{{ GIT_DESCRIBE_NUMBER }} ignore_run_exports_from: {% if cuda_major == "11" %} - {{ compiler('cuda11') }} {% endif %} requirements: build: - {{ compiler('c') }} - {{ compiler('cxx') }} {% if cuda_major == "11" %} - {{ compiler('cuda11') }} ={{ cuda_version }} {% else %} - {{ compiler('cuda') }} {% endif %} - cuda-version ={{ cuda_version }} - cmake {{ cmake_version }} - ninja - sysroot_{{ target_platform }} {{ sysroot_version }} host: - python - libraft {{ version }} - cuda-version ={{ cuda_version }} {% if cuda_major == "11" %} - cuda-profiler-api {{ cuda11_cuda_profiler_api_run_version }} - libcublas {{ cuda11_libcublas_host_version }} - libcublas-dev {{ cuda11_libcublas_host_version }} {% else %} - cuda-profiler-api - libcublas-dev {% endif %} - glog {{ glog_version }} - nlohmann_json {{ nlohmann_json_version }} - h5py {{ h5py_version }} - benchmark - matplotlib - python - pandas - pyyaml # rmm is needed to determine if package is gpu-enabled - rmm ={{ minor_version }} run: - python - libraft {{ version }} - {{ pin_compatible('cuda-version', max_pin='x', min_pin='x') }} {% if cuda_major == "11" %} - cudatoolkit {% endif %} - glog {{ glog_version }} - h5py {{ h5py_version }} - benchmark - glog {{ glog_version }} - matplotlib - python - pandas - pyyaml # rmm is needed to determine if package is gpu-enabled - rmm ={{ minor_version }} about: home: https://rapids.ai/ license: Apache-2.0 summary: RAFT ANN GPU and CPU benchmarks
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/pylibraft/conda_build_config.yaml
c_compiler_version: - 11 cxx_compiler_version: - 11 cuda_compiler: - cuda-nvcc cuda11_compiler: - nvcc sysroot_version: - "2.17" cmake_version: - ">=3.26.4"
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/pylibraft/build.sh
# Copyright (c) 2022, NVIDIA CORPORATION. #!/usr/bin/env bash # This assumes the script is executed from the root of the repo directory ./build.sh pylibraft --no-nvtx
0
rapidsai_public_repos/raft/conda/recipes
rapidsai_public_repos/raft/conda/recipes/pylibraft/meta.yaml
# Copyright (c) 2022-2023, NVIDIA CORPORATION. # Usage: # conda build . -c conda-forge -c numba -c rapidsai -c pytorch {% set version = environ['RAPIDS_PACKAGE_VERSION'].lstrip('v') + environ.get('VERSION_SUFFIX', '') %} {% set minor_version = version.split('.')[0] + '.' + version.split('.')[1] %} {% set py_version = environ['CONDA_PY'] %} {% set cuda_version = '.'.join(environ['RAPIDS_CUDA_VERSION'].split('.')[:2]) %} {% set cuda_major = cuda_version.split('.')[0] %} {% set date_string = environ['RAPIDS_DATE_STRING'] %} package: name: pylibraft version: {{ version }} source: path: ../../.. build: number: {{ GIT_DESCRIBE_NUMBER }} string: cuda{{ cuda_major }}_py{{ py_version }}_{{ date_string }}_{{ GIT_DESCRIBE_HASH }}_{{ GIT_DESCRIBE_NUMBER }} ignore_run_exports_from: {% if cuda_major == "11" %} - {{ compiler('cuda11') }} {% endif %} requirements: build: - {{ compiler('c') }} - {{ compiler('cxx') }} {% if cuda_major == "11" %} - {{ compiler('cuda11') }} ={{ cuda_version }} {% else %} - {{ compiler('cuda') }} {% endif %} - cuda-version ={{ cuda_version }} - cmake {{ cmake_version }} - ninja - sysroot_{{ target_platform }} {{ sysroot_version }} host: {% if cuda_major == "11" %} - cuda-python >=11.7.1,<12.0a0 - cudatoolkit {% else %} - cuda-python >=12.0,<13.0a0 {% endif %} - cuda-version ={{ cuda_version }} - cython >=3.0.0 - libraft {{ version }} - libraft-headers {{ version }} - python x.x - rmm ={{ minor_version }} - scikit-build >=0.13.1 - setuptools run: - {{ pin_compatible('cuda-version', max_pin='x', min_pin='x') }} {% if cuda_major == "11" %} - cudatoolkit {% endif %} - libraft {{ version }} - libraft-headers {{ version }} - numpy >=1.21 - python x.x - rmm ={{ minor_version }} tests: requirements: - cuda-version ={{ cuda_version }} imports: - pylibraft about: home: https://rapids.ai/ license: Apache-2.0 # license_file: LICENSE summary: pylibraft library
0
rapidsai_public_repos/raft
rapidsai_public_repos/raft/cpp/.clangd
# https://clangd.llvm.org/config # Apply a config conditionally to all C files If: PathMatch: .*\.(c|h)$ --- # Apply a config conditionally to all C++ files If: PathMatch: .*\.(c|h)pp --- # Apply a config conditionally to all CUDA files If: PathMatch: .*\.cuh? CompileFlags: Add: - "-x" - "cuda" # No error on unknown CUDA versions - "-Wno-unknown-cuda-version" # Allow variadic CUDA functions - "-Xclang=-fcuda-allow-variadic-functions" Diagnostics: Suppress: - "variadic_device_fn" - "attributes_not_allowed" --- # Tweak the clangd parse settings for all files CompileFlags: Add: # report all errors - "-ferror-limit=0" - "-fmacro-backtrace-limit=0" - "-ftemplate-backtrace-limit=0" # Skip the CUDA version check - "--no-cuda-version-check" Remove: # remove gcc's -fcoroutines - -fcoroutines # remove nvc++ flags unknown to clang - "-gpu=*" - "-stdpar*" # remove nvcc flags unknown to clang - "-arch*" - "-gencode*" - "--generate-code*" - "-ccbin*" - "-t=*" - "--threads*" - "-Xptxas*" - "-Xcudafe*" - "-Xfatbin*" - "-Xcompiler*" - "--diag-suppress*" - "--diag_suppress*" - "--compiler-options*" - "--expt-extended-lambda" - "--expt-relaxed-constexpr" - "-forward-unknown-to-host-compiler" - "-Werror=cross-execution-space-call"
0
rapidsai_public_repos/raft
rapidsai_public_repos/raft/cpp/CMakeLists.txt
# ============================================================================= # Copyright (c) 2020-2023, NVIDIA CORPORATION. # # Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except # in compliance with the License. You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software distributed under the License # is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express # or implied. See the License for the specific language governing permissions and limitations under # the License. set(RAPIDS_VERSION "24.02") set(RAFT_VERSION "24.02.00") cmake_minimum_required(VERSION 3.26.4 FATAL_ERROR) include(../fetch_rapids.cmake) include(rapids-cmake) include(rapids-cpm) include(rapids-export) include(rapids-find) option(BUILD_CPU_ONLY "Build CPU only components. Applies to RAFT ANN benchmarks currently" OFF) # workaround for rapids_cuda_init_architectures not working for arch detection with # enable_language(CUDA) set(lang_list "CXX") if(NOT BUILD_CPU_ONLY) include(rapids-cuda) rapids_cuda_init_architectures(RAFT) list(APPEND lang_list "CUDA") endif() project( RAFT VERSION ${RAFT_VERSION} LANGUAGES ${lang_list} ) # Write the version header rapids_cmake_write_version_file(include/raft/version_config.hpp) # ################################################################################################## # * build type --------------------------------------------------------------- # Set a default build type if none was specified rapids_cmake_build_type(Release) # this is needed for clang-tidy runs set(CMAKE_EXPORT_COMPILE_COMMANDS ON) # ################################################################################################## # * User Options ------------------------------------------------------------ option(BUILD_SHARED_LIBS "Build raft shared libraries" ON) option(BUILD_TESTS "Build raft unit-tests" ON) option(BUILD_PRIMS_BENCH "Build raft C++ benchmark tests" OFF) option(BUILD_ANN_BENCH "Build raft ann benchmarks" OFF) option(CUDA_ENABLE_KERNELINFO "Enable kernel resource usage info" OFF) option(CUDA_ENABLE_LINEINFO "Enable the -lineinfo option for nvcc (useful for cuda-memcheck / profiler)" OFF ) option(CUDA_STATIC_RUNTIME "Statically link the CUDA toolkit runtime and libraries" OFF) option(CUDA_LOG_COMPILE_TIME "Write a log of compilation times to nvcc_compile_log.csv" OFF) option(DETECT_CONDA_ENV "Enable detection of conda environment for dependencies" ON) option(DISABLE_DEPRECATION_WARNINGS "Disable deprecaction warnings " ON) option(DISABLE_OPENMP "Disable OpenMP" OFF) option(RAFT_NVTX "Enable nvtx markers" OFF) set(RAFT_COMPILE_LIBRARY_DEFAULT OFF) if((BUILD_TESTS OR BUILD_PRIMS_BENCH OR BUILD_ANN_BENCH ) AND NOT BUILD_CPU_ONLY ) set(RAFT_COMPILE_LIBRARY_DEFAULT ON) endif() option(RAFT_COMPILE_LIBRARY "Enable building raft shared library instantiations" ${RAFT_COMPILE_LIBRARY_DEFAULT} ) if(BUILD_CPU_ONLY) set(BUILD_SHARED_LIBS OFF) set(BUILD_TESTS OFF) endif() # Needed because GoogleBenchmark changes the state of FindThreads.cmake, causing subsequent runs to # have different values for the `Threads::Threads` target. Setting this flag ensures # `Threads::Threads` is the same value across all builds so that cache hits occur set(THREADS_PREFER_PTHREAD_FLAG ON) include(CMakeDependentOption) # cmake_dependent_option( RAFT_USE_FAISS_STATIC "Build and statically link the FAISS library for # nearest neighbors search on GPU" ON RAFT_COMPILE_LIBRARY OFF ) message(VERBOSE "RAFT: Building optional components: ${raft_FIND_COMPONENTS}") message(VERBOSE "RAFT: Build RAFT unit-tests: ${BUILD_TESTS}") message(VERBOSE "RAFT: Building raft C++ benchmarks: ${BUILD_PRIMS_BENCH}") message(VERBOSE "RAFT: Building ANN benchmarks: ${BUILD_ANN_BENCH}") message(VERBOSE "RAFT: Build CPU only components: ${BUILD_CPU_ONLY}") message(VERBOSE "RAFT: Enable detection of conda environment for dependencies: ${DETECT_CONDA_ENV}") message(VERBOSE "RAFT: Disable depreaction warnings " ${DISABLE_DEPRECATION_WARNINGS}) message(VERBOSE "RAFT: Disable OpenMP: ${DISABLE_OPENMP}") message(VERBOSE "RAFT: Enable kernel resource usage info: ${CUDA_ENABLE_KERNELINFO}") message(VERBOSE "RAFT: Enable lineinfo in nvcc: ${CUDA_ENABLE_LINEINFO}") message(VERBOSE "RAFT: Enable nvtx markers: ${RAFT_NVTX}") message(VERBOSE "RAFT: Statically link the CUDA toolkit runtime and libraries: ${CUDA_STATIC_RUNTIME}" ) # Set RMM logging level set(RMM_LOGGING_LEVEL "INFO" CACHE STRING "Choose the logging level." ) set_property( CACHE RMM_LOGGING_LEVEL PROPERTY STRINGS "TRACE" "DEBUG" "INFO" "WARN" "ERROR" "CRITICAL" "OFF" ) message(VERBOSE "RAFT: RMM_LOGGING_LEVEL = '${RMM_LOGGING_LEVEL}'.") # ################################################################################################## # * Conda environment detection ---------------------------------------------- if(DETECT_CONDA_ENV) rapids_cmake_support_conda_env(conda_env MODIFY_PREFIX_PATH) if(CMAKE_INSTALL_PREFIX_INITIALIZED_TO_DEFAULT AND DEFINED ENV{CONDA_PREFIX}) message( STATUS "RAFT: No CMAKE_INSTALL_PREFIX argument detected, setting to: $ENV{CONDA_PREFIX}" ) set(CMAKE_INSTALL_PREFIX "$ENV{CONDA_PREFIX}") endif() endif() # ################################################################################################## # * compiler options ---------------------------------------------------------- set(_ctk_static_suffix "") if(CUDA_STATIC_RUNTIME) set(_ctk_static_suffix "_static") endif() if(NOT BUILD_CPU_ONLY) # CUDA runtime rapids_cuda_init_runtime(USE_STATIC ${CUDA_STATIC_RUNTIME}) # * find CUDAToolkit package # * determine GPU architectures # * enable the CMake CUDA language # * set other CUDA compilation flags rapids_find_package( CUDAToolkit REQUIRED BUILD_EXPORT_SET raft-exports INSTALL_EXPORT_SET raft-exports ) else() add_compile_definitions(BUILD_CPU_ONLY) endif() if(NOT DISABLE_OPENMP) rapids_find_package( OpenMP REQUIRED BUILD_EXPORT_SET raft-exports INSTALL_EXPORT_SET raft-exports ) if(OPENMP_FOUND) message(VERBOSE "RAFT: OpenMP found in ${OpenMP_CXX_INCLUDE_DIRS}") endif() endif() include(cmake/modules/ConfigureCUDA.cmake) # ################################################################################################## # * Requirements ------------------------------------------------------------- # add third party dependencies using CPM rapids_cpm_init() if(NOT BUILD_CPU_ONLY) # thrust before rmm/cuco so we get the right version of thrust/cub include(cmake/thirdparty/get_thrust.cmake) include(cmake/thirdparty/get_rmm.cmake) include(cmake/thirdparty/get_cutlass.cmake) include(${rapids-cmake-dir}/cpm/cuco.cmake) rapids_cpm_cuco(BUILD_EXPORT_SET raft-exports INSTALL_EXPORT_SET raft-exports) endif() if(BUILD_TESTS) include(cmake/thirdparty/get_gtest.cmake) endif() if(BUILD_PRIMS_BENCH OR BUILD_ANN_BENCH) include(${rapids-cmake-dir}/cpm/gbench.cmake) rapids_cpm_gbench() endif() # ################################################################################################## # * raft --------------------------------------------------------------------- add_library(raft INTERFACE) add_library(raft::raft ALIAS raft) target_include_directories( raft INTERFACE "$<BUILD_INTERFACE:${RAFT_SOURCE_DIR}/include>" "$<INSTALL_INTERFACE:include>" ) if(NOT BUILD_CPU_ONLY) # Keep RAFT as lightweight as possible. Only CUDA libs and rmm should be used in global target. target_link_libraries(raft INTERFACE rmm::rmm cuco::cuco nvidia::cutlass::cutlass raft::Thrust) endif() target_compile_features(raft INTERFACE cxx_std_17 $<BUILD_INTERFACE:cuda_std_17>) target_compile_options( raft INTERFACE $<$<COMPILE_LANG_AND_ID:CUDA,NVIDIA>:--expt-extended-lambda --expt-relaxed-constexpr> ) set(RAFT_CUSOLVER_DEPENDENCY CUDA::cusolver${_ctk_static_suffix}) set(RAFT_CUBLAS_DEPENDENCY CUDA::cublas${_ctk_static_suffix}) set(RAFT_CURAND_DEPENDENCY CUDA::curand${_ctk_static_suffix}) set(RAFT_CUSPARSE_DEPENDENCY CUDA::cusparse${_ctk_static_suffix}) set(RAFT_CTK_MATH_DEPENDENCIES ${RAFT_CUBLAS_DEPENDENCY} ${RAFT_CUSOLVER_DEPENDENCY} ${RAFT_CUSPARSE_DEPENDENCY} ${RAFT_CURAND_DEPENDENCY} ) # Endian detection include(TestBigEndian) test_big_endian(BIG_ENDIAN) if(BIG_ENDIAN) target_compile_definitions(raft INTERFACE RAFT_SYSTEM_LITTLE_ENDIAN=0) else() target_compile_definitions(raft INTERFACE RAFT_SYSTEM_LITTLE_ENDIAN=1) endif() if(RAFT_COMPILE_LIBRARY) file( WRITE "${CMAKE_CURRENT_BINARY_DIR}/fatbin.ld" [=[ SECTIONS { .nvFatBinSegment : { *(.nvFatBinSegment) } .nv_fatbin : { *(.nv_fatbin) } } ]=] ) endif() # ################################################################################################## # * NVTX support in raft ----------------------------------------------------- if(RAFT_NVTX) # This enables NVTX within the project with no option to disable it downstream. target_link_libraries(raft INTERFACE CUDA::nvToolsExt) target_compile_definitions(raft INTERFACE NVTX_ENABLED) else() # Allow enable NVTX downstream if not set here. This creates a new option at build/install time, # which is set by default to OFF, but can be enabled in the dependent project. get_property( nvtx_option_help_string CACHE RAFT_NVTX PROPERTY HELPSTRING ) string( CONCAT nvtx_export_string "option(RAFT_NVTX \"" ${nvtx_option_help_string} "\" OFF)" [=[ target_link_libraries(raft::raft INTERFACE $<$<BOOL:${RAFT_NVTX}>:CUDA::nvToolsExt>) target_compile_definitions(raft::raft INTERFACE $<$<BOOL:${RAFT_NVTX}>:NVTX_ENABLED>) ]=] ) endif() # ################################################################################################## # * raft_compiled ------------------------------------------------------------ add_library(raft_compiled INTERFACE) if(TARGET raft_compiled AND (NOT TARGET raft::compiled)) add_library(raft::compiled ALIAS raft_compiled) endif() set_target_properties(raft_compiled PROPERTIES EXPORT_NAME compiled) if(RAFT_COMPILE_LIBRARY) add_library( raft_objs OBJECT src/core/logger.cpp src/distance/detail/pairwise_matrix/dispatch_canberra_double_double_double_int.cu src/distance/detail/pairwise_matrix/dispatch_canberra_float_float_float_int.cu src/distance/detail/pairwise_matrix/dispatch_correlation_double_double_double_int.cu src/distance/detail/pairwise_matrix/dispatch_correlation_float_float_float_int.cu src/distance/detail/pairwise_matrix/dispatch_cosine_double_double_double_int.cu src/distance/detail/pairwise_matrix/dispatch_cosine_float_float_float_int.cu src/distance/detail/pairwise_matrix/dispatch_hamming_unexpanded_double_double_double_int.cu src/distance/detail/pairwise_matrix/dispatch_hamming_unexpanded_float_float_float_int.cu src/distance/detail/pairwise_matrix/dispatch_hellinger_expanded_double_double_double_int.cu src/distance/detail/pairwise_matrix/dispatch_hellinger_expanded_float_float_float_int.cu src/distance/detail/pairwise_matrix/dispatch_jensen_shannon_double_double_double_int.cu src/distance/detail/pairwise_matrix/dispatch_jensen_shannon_float_float_float_int.cu src/distance/detail/pairwise_matrix/dispatch_kl_divergence_double_double_double_int.cu src/distance/detail/pairwise_matrix/dispatch_kl_divergence_float_float_float_int.cu src/distance/detail/pairwise_matrix/dispatch_l1_double_double_double_int.cu src/distance/detail/pairwise_matrix/dispatch_l1_float_float_float_int.cu src/distance/detail/pairwise_matrix/dispatch_l2_expanded_double_double_double_int.cu src/distance/detail/pairwise_matrix/dispatch_l2_expanded_float_float_float_int.cu src/distance/detail/pairwise_matrix/dispatch_l2_unexpanded_double_double_double_int.cu src/distance/detail/pairwise_matrix/dispatch_l2_unexpanded_float_float_float_int.cu src/distance/detail/pairwise_matrix/dispatch_l_inf_double_double_double_int.cu src/distance/detail/pairwise_matrix/dispatch_l_inf_float_float_float_int.cu src/distance/detail/pairwise_matrix/dispatch_lp_unexpanded_double_double_double_int.cu src/distance/detail/pairwise_matrix/dispatch_lp_unexpanded_float_float_float_int.cu src/distance/detail/pairwise_matrix/dispatch_rbf.cu src/distance/detail/pairwise_matrix/dispatch_russel_rao_double_double_double_int.cu src/distance/detail/pairwise_matrix/dispatch_russel_rao_float_float_float_int.cu src/distance/distance.cu src/distance/fused_l2_nn.cu src/linalg/detail/coalesced_reduction.cu src/matrix/detail/select_k_double_int64_t.cu src/matrix/detail/select_k_double_uint32_t.cu src/matrix/detail/select_k_float_int64_t.cu src/matrix/detail/select_k_float_uint32_t.cu src/matrix/detail/select_k_float_int32.cu src/matrix/detail/select_k_half_int64_t.cu src/matrix/detail/select_k_half_uint32_t.cu src/neighbors/ball_cover.cu src/neighbors/brute_force_fused_l2_knn_float_int64_t.cu src/neighbors/brute_force_knn_int64_t_float_int64_t.cu src/neighbors/brute_force_knn_int64_t_float_uint32_t.cu src/neighbors/brute_force_knn_int_float_int.cu src/neighbors/brute_force_knn_uint32_t_float_uint32_t.cu src/neighbors/brute_force_knn_index_float.cu src/neighbors/detail/cagra/search_multi_cta_float_uint32_dim128_t8.cu src/neighbors/detail/cagra/search_multi_cta_float_uint32_dim256_t16.cu src/neighbors/detail/cagra/search_multi_cta_float_uint32_dim512_t32.cu src/neighbors/detail/cagra/search_multi_cta_float_uint32_dim1024_t32.cu src/neighbors/detail/cagra/search_multi_cta_int8_uint32_dim128_t8.cu src/neighbors/detail/cagra/search_multi_cta_int8_uint32_dim256_t16.cu src/neighbors/detail/cagra/search_multi_cta_int8_uint32_dim512_t32.cu src/neighbors/detail/cagra/search_multi_cta_int8_uint32_dim1024_t32.cu src/neighbors/detail/cagra/search_multi_cta_uint8_uint32_dim128_t8.cu src/neighbors/detail/cagra/search_multi_cta_uint8_uint32_dim256_t16.cu src/neighbors/detail/cagra/search_multi_cta_uint8_uint32_dim512_t32.cu src/neighbors/detail/cagra/search_multi_cta_uint8_uint32_dim1024_t32.cu src/neighbors/detail/cagra/search_single_cta_float_uint32_dim128_t8.cu src/neighbors/detail/cagra/search_single_cta_float_uint32_dim256_t16.cu src/neighbors/detail/cagra/search_single_cta_float_uint32_dim512_t32.cu src/neighbors/detail/cagra/search_single_cta_float_uint32_dim1024_t32.cu src/neighbors/detail/cagra/search_single_cta_int8_uint32_dim128_t8.cu src/neighbors/detail/cagra/search_single_cta_int8_uint32_dim256_t16.cu src/neighbors/detail/cagra/search_single_cta_int8_uint32_dim512_t32.cu src/neighbors/detail/cagra/search_single_cta_int8_uint32_dim1024_t32.cu src/neighbors/detail/cagra/search_single_cta_uint8_uint32_dim128_t8.cu src/neighbors/detail/cagra/search_single_cta_uint8_uint32_dim256_t16.cu src/neighbors/detail/cagra/search_single_cta_uint8_uint32_dim512_t32.cu src/neighbors/detail/cagra/search_single_cta_uint8_uint32_dim1024_t32.cu src/neighbors/detail/ivf_flat_interleaved_scan_float_float_int64_t.cu src/neighbors/detail/ivf_flat_interleaved_scan_int8_t_int32_t_int64_t.cu src/neighbors/detail/ivf_flat_interleaved_scan_uint8_t_uint32_t_int64_t.cu src/neighbors/detail/ivf_flat_search.cu src/neighbors/detail/ivf_pq_compute_similarity_float_float.cu src/neighbors/detail/ivf_pq_compute_similarity_float_fp8_false.cu src/neighbors/detail/ivf_pq_compute_similarity_float_fp8_true.cu src/neighbors/detail/ivf_pq_compute_similarity_float_half.cu src/neighbors/detail/ivf_pq_compute_similarity_half_fp8_false.cu src/neighbors/detail/ivf_pq_compute_similarity_half_fp8_true.cu src/neighbors/detail/ivf_pq_compute_similarity_half_half.cu src/neighbors/detail/refine_host_float_float.cpp src/neighbors/detail/refine_host_int8_t_float.cpp src/neighbors/detail/refine_host_uint8_t_float.cpp src/neighbors/detail/selection_faiss_int32_t_float.cu src/neighbors/detail/selection_faiss_int_double.cu src/neighbors/detail/selection_faiss_long_float.cu src/neighbors/detail/selection_faiss_size_t_double.cu src/neighbors/detail/selection_faiss_size_t_float.cu src/neighbors/detail/selection_faiss_uint32_t_float.cu src/neighbors/detail/selection_faiss_int64_t_double.cu src/neighbors/detail/selection_faiss_int64_t_half.cu src/neighbors/detail/selection_faiss_uint32_t_double.cu src/neighbors/detail/selection_faiss_uint32_t_half.cu src/neighbors/ivf_flat_build_float_int64_t.cu src/neighbors/ivf_flat_build_int8_t_int64_t.cu src/neighbors/ivf_flat_build_uint8_t_int64_t.cu src/neighbors/ivf_flat_extend_float_int64_t.cu src/neighbors/ivf_flat_extend_int8_t_int64_t.cu src/neighbors/ivf_flat_extend_uint8_t_int64_t.cu src/neighbors/ivf_flat_search_float_int64_t.cu src/neighbors/ivf_flat_search_int8_t_int64_t.cu src/neighbors/ivf_flat_search_uint8_t_int64_t.cu src/neighbors/ivfpq_build_float_int64_t.cu src/neighbors/ivfpq_build_int8_t_int64_t.cu src/neighbors/ivfpq_build_uint8_t_int64_t.cu src/neighbors/ivfpq_extend_float_int64_t.cu src/neighbors/ivfpq_extend_int8_t_int64_t.cu src/neighbors/ivfpq_extend_uint8_t_int64_t.cu src/neighbors/ivfpq_search_float_int64_t.cu src/neighbors/ivfpq_search_int8_t_int64_t.cu src/neighbors/ivfpq_search_uint8_t_int64_t.cu src/neighbors/refine_float_float.cu src/neighbors/refine_int8_t_float.cu src/neighbors/refine_uint8_t_float.cu src/raft_runtime/cluster/cluster_cost.cuh src/raft_runtime/cluster/cluster_cost_double.cu src/raft_runtime/cluster/cluster_cost_float.cu src/raft_runtime/cluster/kmeans_fit_double.cu src/raft_runtime/cluster/kmeans_fit_float.cu src/raft_runtime/cluster/kmeans_init_plus_plus_double.cu src/raft_runtime/cluster/kmeans_init_plus_plus_float.cu src/raft_runtime/cluster/update_centroids.cuh src/raft_runtime/cluster/update_centroids_double.cu src/raft_runtime/cluster/update_centroids_float.cu src/raft_runtime/distance/fused_l2_min_arg.cu src/raft_runtime/distance/pairwise_distance.cu src/raft_runtime/matrix/select_k_float_int64_t.cu src/raft_runtime/neighbors/brute_force_knn_int64_t_float.cu src/raft_runtime/neighbors/cagra_build.cu src/raft_runtime/neighbors/cagra_search.cu src/raft_runtime/neighbors/cagra_serialize.cu src/raft_runtime/neighbors/ivf_flat_build.cu src/raft_runtime/neighbors/ivf_flat_search.cu src/raft_runtime/neighbors/ivf_flat_serialize.cu src/raft_runtime/neighbors/ivfpq_build.cu src/raft_runtime/neighbors/ivfpq_deserialize.cu src/raft_runtime/neighbors/ivfpq_search_float_int64_t.cu src/raft_runtime/neighbors/ivfpq_search_int8_t_int64_t.cu src/raft_runtime/neighbors/ivfpq_search_uint8_t_int64_t.cu src/raft_runtime/neighbors/ivfpq_serialize.cu src/raft_runtime/neighbors/refine_d_int64_t_float.cu src/raft_runtime/neighbors/refine_d_int64_t_int8_t.cu src/raft_runtime/neighbors/refine_d_int64_t_uint8_t.cu src/raft_runtime/neighbors/refine_h_int64_t_float.cu src/raft_runtime/neighbors/refine_h_int64_t_int8_t.cu src/raft_runtime/neighbors/refine_h_int64_t_uint8_t.cu src/raft_runtime/random/rmat_rectangular_generator_int64_double.cu src/raft_runtime/random/rmat_rectangular_generator_int64_float.cu src/raft_runtime/random/rmat_rectangular_generator_int_double.cu src/raft_runtime/random/rmat_rectangular_generator_int_float.cu src/spatial/knn/detail/ball_cover/registers_pass_one_2d_dist.cu src/spatial/knn/detail/ball_cover/registers_pass_one_2d_euclidean.cu src/spatial/knn/detail/ball_cover/registers_pass_one_2d_haversine.cu src/spatial/knn/detail/ball_cover/registers_pass_one_3d_dist.cu src/spatial/knn/detail/ball_cover/registers_pass_one_3d_euclidean.cu src/spatial/knn/detail/ball_cover/registers_pass_one_3d_haversine.cu src/spatial/knn/detail/ball_cover/registers_pass_two_2d_dist.cu src/spatial/knn/detail/ball_cover/registers_pass_two_2d_euclidean.cu src/spatial/knn/detail/ball_cover/registers_pass_two_2d_haversine.cu src/spatial/knn/detail/ball_cover/registers_pass_two_3d_dist.cu src/spatial/knn/detail/ball_cover/registers_pass_two_3d_euclidean.cu src/spatial/knn/detail/ball_cover/registers_pass_two_3d_haversine.cu src/spatial/knn/detail/fused_l2_knn_int32_t_float.cu src/spatial/knn/detail/fused_l2_knn_int64_t_float.cu src/spatial/knn/detail/fused_l2_knn_uint32_t_float.cu src/util/memory_pool.cpp ) set_target_properties( raft_objs PROPERTIES CXX_STANDARD 17 CXX_STANDARD_REQUIRED ON CUDA_STANDARD 17 CUDA_STANDARD_REQUIRED ON POSITION_INDEPENDENT_CODE ON ) target_compile_definitions(raft_objs PRIVATE "RAFT_EXPLICIT_INSTANTIATE_ONLY") target_compile_options( raft_objs PRIVATE "$<$<COMPILE_LANGUAGE:CXX>:${RAFT_CXX_FLAGS}>" "$<$<COMPILE_LANGUAGE:CUDA>:${RAFT_CUDA_FLAGS}>" ) add_library(raft_lib SHARED $<TARGET_OBJECTS:raft_objs>) add_library(raft_lib_static STATIC $<TARGET_OBJECTS:raft_objs>) set_target_properties( raft_lib raft_lib_static PROPERTIES OUTPUT_NAME raft BUILD_RPATH "\$ORIGIN" INSTALL_RPATH "\$ORIGIN" INTERFACE_POSITION_INDEPENDENT_CODE ON ) foreach(target raft_lib raft_lib_static raft_objs) target_link_libraries( ${target} PUBLIC raft::raft ${RAFT_CTK_MATH_DEPENDENCIES} # TODO: Once `raft::resources` is used everywhere, this # will just be cublas $<TARGET_NAME_IF_EXISTS:OpenMP::OpenMP_CXX> ) # So consumers know when using libraft.so/libraft.a target_compile_definitions(${target} PUBLIC "RAFT_COMPILED") # ensure CUDA symbols aren't relocated to the middle of the debug build binaries target_link_options(${target} PRIVATE "${CMAKE_CURRENT_BINARY_DIR}/fatbin.ld") endforeach() endif() if(TARGET raft_lib AND (NOT TARGET raft::raft_lib)) add_library(raft::raft_lib ALIAS raft_lib) endif() target_link_libraries(raft_compiled INTERFACE raft::raft $<TARGET_NAME_IF_EXISTS:raft::raft_lib>) # ################################################################################################## # * raft_compiled_static---------------------------------------------------------------------------- add_library(raft_compiled_static INTERFACE) if(TARGET raft_compiled_static AND (NOT TARGET raft::compiled_static)) add_library(raft::compiled_static ALIAS raft_compiled_static) endif() set_target_properties(raft_compiled_static PROPERTIES EXPORT_NAME compiled_static) if(TARGET raft_lib_static AND (NOT TARGET raft::raft_lib_static)) add_library(raft::raft_lib_static ALIAS raft_lib_static) endif() target_link_libraries( raft_compiled_static INTERFACE raft::raft $<TARGET_NAME_IF_EXISTS:raft::raft_lib_static> ) # ################################################################################################## # * raft_distributed ------------------------------------------------------------------------------- add_library(raft_distributed INTERFACE) if(TARGET raft_distributed AND (NOT TARGET raft::distributed)) add_library(raft::distributed ALIAS raft_distributed) endif() set_target_properties(raft_distributed PROPERTIES EXPORT_NAME distributed) rapids_find_generate_module( NCCL HEADER_NAMES nccl.h LIBRARY_NAMES nccl BUILD_EXPORT_SET raft-distributed-exports INSTALL_EXPORT_SET raft-distributed-exports ) rapids_export_package(BUILD ucx raft-distributed-exports) rapids_export_package(INSTALL ucx raft-distributed-exports) rapids_export_package(BUILD NCCL raft-distributed-exports) rapids_export_package(INSTALL NCCL raft-distributed-exports) target_link_libraries(raft_distributed INTERFACE ucx::ucp NCCL::NCCL) # ################################################################################################## # * install targets----------------------------------------------------------- rapids_cmake_install_lib_dir(lib_dir) include(GNUInstallDirs) include(CPack) install( TARGETS raft DESTINATION ${lib_dir} COMPONENT raft EXPORT raft-exports ) install( TARGETS raft_compiled raft_compiled_static DESTINATION ${lib_dir} COMPONENT raft EXPORT raft-compiled-exports ) if(TARGET raft_lib) install( TARGETS raft_lib DESTINATION ${lib_dir} COMPONENT compiled EXPORT raft-compiled-lib-exports ) install( TARGETS raft_lib_static DESTINATION ${lib_dir} COMPONENT compiled-static EXPORT raft-compiled-static-lib-exports ) install( DIRECTORY include/raft_runtime DESTINATION ${CMAKE_INSTALL_INCLUDEDIR} COMPONENT compiled ) endif() install( TARGETS raft_distributed DESTINATION ${lib_dir} COMPONENT distributed EXPORT raft-distributed-exports ) install( DIRECTORY include/raft COMPONENT raft DESTINATION ${CMAKE_INSTALL_INCLUDEDIR} ) # Temporary install of raft.hpp while the file is removed install( FILES include/raft.hpp COMPONENT raft DESTINATION ${CMAKE_INSTALL_INCLUDEDIR}/raft ) install( FILES ${CMAKE_CURRENT_BINARY_DIR}/include/raft/version_config.hpp COMPONENT raft DESTINATION include/raft ) # ################################################################################################## # * install export ----------------------------------------------------------- set(doc_string [=[ Provide targets for the RAFT: Reusable Accelerated Functions and Tools RAFT contains fundamental widely-used algorithms and primitives for data science and machine learning. Optional Components: - compiled - compiled_static - distributed Imported Targets: - raft::raft - raft::compiled brought in by the `compiled` optional component - raft::compiled_static brought in by the `compiled_static` optional component - raft::distributed brought in by the `distributed` optional component ]=] ) set(code_string ${nvtx_export_string}) string( APPEND code_string [=[ if(NOT TARGET raft::Thrust) thrust_create_target(raft::Thrust FROM_OPTIONS) endif() ]=] ) string( APPEND code_string [=[ if(compiled IN_LIST raft_FIND_COMPONENTS) enable_language(CUDA) endif() ]=] ) set(raft_components compiled distributed) set(raft_export_sets raft-compiled-exports raft-distributed-exports) if(TARGET raft_lib) list(APPEND raft_components compiled compiled-static) list(APPEND raft_export_sets raft-compiled-lib-exports raft-compiled-static-lib-exports) endif() string( APPEND code_string [=[ option(RAFT_ENABLE_CUSOLVER_DEPENDENCY "Enable cusolver dependency" ON) option(RAFT_ENABLE_CUBLAS_DEPENDENCY "Enable cublas dependency" ON) option(RAFT_ENABLE_CURAND_DEPENDENCY "Enable curand dependency" ON) option(RAFT_ENABLE_CUSPARSE_DEPENDENCY "Enable cusparse dependency" ON) mark_as_advanced(RAFT_ENABLE_CUSOLVER_DEPENDENCY) mark_as_advanced(RAFT_ENABLE_CUBLAS_DEPENDENCY) mark_as_advanced(RAFT_ENABLE_CURAND_DEPENDENCY) mark_as_advanced(RAFT_ENABLE_CUSPARSE_DEPENDENCY) target_link_libraries(raft::raft INTERFACE $<$<BOOL:${RAFT_ENABLE_CUSOLVER_DEPENDENCY}>:${RAFT_CUSOLVER_DEPENDENCY}> $<$<BOOL:${RAFT_ENABLE_CUBLAS_DEPENDENCY}>:${RAFT_CUBLAS_DEPENDENCY}> $<$<BOOL:${RAFT_ENABLE_CUSPARSE_DEPENDENCY}>:${RAFT_CUSPARSE_DEPENDENCY}> $<$<BOOL:${RAFT_ENABLE_CURAND_DEPENDENCY}>:${RAFT_CURAND_DEPENDENCY}> ) ]=] ) # Use `rapids_export` for 22.04 as it will have COMPONENT support rapids_export( INSTALL raft EXPORT_SET raft-exports COMPONENTS ${raft_components} COMPONENTS_EXPORT_SET ${raft_export_sets} GLOBAL_TARGETS raft compiled distributed NAMESPACE raft:: DOCUMENTATION doc_string FINAL_CODE_BLOCK code_string ) # ################################################################################################## # * build export ------------------------------------------------------------- rapids_export( BUILD raft EXPORT_SET raft-exports COMPONENTS ${raft_components} COMPONENTS_EXPORT_SET ${raft_export_sets} GLOBAL_TARGETS raft compiled distributed DOCUMENTATION doc_string NAMESPACE raft:: FINAL_CODE_BLOCK code_string ) # ################################################################################################## # * shared test/bench headers ------------------------------------------------ if(BUILD_TESTS OR BUILD_PRIMS_BENCH) include(internal/CMakeLists.txt) endif() # ################################################################################################## # * build test executable ---------------------------------------------------- if(BUILD_TESTS) include(test/CMakeLists.txt) endif() # ################################################################################################## # * build benchmark executable ----------------------------------------------- if(BUILD_PRIMS_BENCH) include(bench/prims/CMakeLists.txt) endif() # ################################################################################################## # * build ann benchmark executable ----------------------------------------------- if(BUILD_ANN_BENCH) include(bench/ann/CMakeLists.txt) endif()
0
rapidsai_public_repos/raft
rapidsai_public_repos/raft/cpp/.clang-tidy
--- # Refer to the following link for the explanation of each params: # https://releases.llvm.org/8.0.1/tools/clang/tools/extra/docs/clang-tidy/checks/list.html Checks: 'clang-diagnostic-*,clang-analyzer-*,modernize-*,-modernize-make-*,-modernize-raw-string-literal,google-*,-google-default-arguments,-clang-diagnostic-#pragma-messages,readability-identifier-naming,-*,modernize-*,-modernize-make-*,-modernize-raw-string-literal,google-*,-google-default-arguments,-clang-diagnostic-#pragma-messages,readability-identifier-naming' WarningsAsErrors: '' HeaderFilterRegex: '' AnalyzeTemporaryDtors: false FormatStyle: none User: snanditale CheckOptions: - key: google-build-namespaces.HeaderFileExtensions value: ',h,hh,hpp,hxx' - key: google-global-names-in-headers.HeaderFileExtensions value: ',h,hh,hpp,hxx' - key: google-readability-braces-around-statements.ShortStatementLines value: '1' - key: google-readability-function-size.BranchThreshold value: '4294967295' - key: google-readability-function-size.LineThreshold value: '4294967295' - key: google-readability-function-size.NestingThreshold value: '4294967295' - key: google-readability-function-size.ParameterThreshold value: '4294967295' - key: google-readability-function-size.StatementThreshold value: '800' - key: google-readability-function-size.VariableThreshold value: '4294967295' - key: google-readability-namespace-comments.ShortNamespaceLines value: '10' - key: google-readability-namespace-comments.SpacesBeforeComments value: '2' - key: google-runtime-int.SignedTypePrefix value: int - key: google-runtime-int.TypeSuffix value: '' - key: google-runtime-int.UnsignedTypePrefix value: uint - key: google-runtime-references.WhiteListTypes value: '' - key: modernize-loop-convert.MaxCopySize value: '16' - key: modernize-loop-convert.MinConfidence value: reasonable - key: modernize-loop-convert.NamingStyle value: CamelCase - key: modernize-pass-by-value.IncludeStyle value: llvm - key: modernize-pass-by-value.ValuesOnly value: '0' - key: modernize-replace-auto-ptr.IncludeStyle value: llvm - key: modernize-replace-random-shuffle.IncludeStyle value: llvm - key: modernize-use-auto.MinTypeNameLength value: '5' - key: modernize-use-auto.RemoveStars value: '0' - key: modernize-use-default-member-init.IgnoreMacros value: '1' - key: modernize-use-default-member-init.UseAssignment value: '0' - key: modernize-use-emplace.ContainersWithPushBack value: '::std::vector;::std::list;::std::deque' - key: modernize-use-emplace.SmartPointers value: '::std::shared_ptr;::std::unique_ptr;::std::auto_ptr;::std::weak_ptr' - key: modernize-use-emplace.TupleMakeFunctions value: '::std::make_pair;::std::make_tuple' - key: modernize-use-emplace.TupleTypes value: '::std::pair;::std::tuple' - key: modernize-use-equals-default.IgnoreMacros value: '1' - key: modernize-use-noexcept.ReplacementString value: '' - key: modernize-use-noexcept.UseNoexceptFalse value: '1' - key: modernize-use-nullptr.NullMacros value: 'NULL' - key: modernize-use-transparent-functors.SafeMode value: '0' - key: modernize-use-using.IgnoreMacros value: '1' - key: readability-identifier-naming.AbstractClassCase value: lower_case - key: readability-identifier-naming.AbstractClassPrefix value: '' - key: readability-identifier-naming.AbstractClassSuffix value: '' - key: readability-identifier-naming.ClassCase value: lower_case - key: readability-identifier-naming.ClassPrefix value: '' - key: readability-identifier-naming.ClassSuffix value: '' - key: readability-identifier-naming.ClassConstantCase value: CamelCase - key: readability-identifier-naming.ClassConstantPrefix value: 'k' - key: readability-identifier-naming.ClassConstantSuffix value: '' - key: readability-identifier-naming.ClassMemberCase value: lower_case - key: readability-identifier-naming.ClassMemberPrefix value: '' - key: readability-identifier-naming.ClassMemberSuffix value: '_' - key: readability-identifier-naming.ClassMethodCase value: lower_case - key: readability-identifier-naming.ClassMethodPrefix value: '' - key: readability-identifier-naming.ClassMethodSuffix value: '' - key: readability-identifier-naming.ConstexprFunctionCase value: lower_case - key: readability-identifier-naming.ConstexprFunctionPrefix value: '' - key: readability-identifier-naming.ConstexprFunctionSuffix value: '' - key: readability-identifier-naming.ConstexprMethodCase value: lower_case - key: readability-identifier-naming.ConstexprMethodPrefix value: '' - key: readability-identifier-naming.ConstexprMethodSuffix value: '' - key: readability-identifier-naming.ConstexprVariableCase value: CamelCase - key: readability-identifier-naming.ConstexprVariablePrefix value: 'k' - key: readability-identifier-naming.ConstexprVariableSuffix value: '' - key: readability-identifier-naming.EnumCase value: CamelCase - key: readability-identifier-naming.EnumPrefix value: '' - key: readability-identifier-naming.EnumSuffix value: '' - key: readability-identifier-naming.EnumConstantCase value: CamelCase - key: readability-identifier-naming.EnumConstantPrefix value: 'k' - key: readability-identifier-naming.EnumConstantSuffix value: '' - key: readability-identifier-naming.FunctionCase value: lower_case - key: readability-identifier-naming.FunctionPrefix value: '' - key: readability-identifier-naming.FunctionSuffix value: '' - key: readability-identifier-naming.GlobalConstantCase value: CamelCase - key: readability-identifier-naming.GlobalConstantPrefix value: 'k' - key: readability-identifier-naming.GlobalConstantSuffix value: '' - key: readability-identifier-naming.IgnoreFailedSplit value: '0' - key: readability-identifier-naming.LocalVariableCase value: 'lower_case' - key: readability-identifier-naming.LocalVariablePrefix value: '' - key: readability-identifier-naming.LocalVariableSuffix value: '' - key: readability-identifier-naming.ConstExprVariableCase value: 'CamelCase' - key: readability-identifier-naming.ConstExprVariablePrefix value: 'k' - key: readability-identifier-naming.ConstExprVariableSuffix value: '' - key: readability-identifier-naming.MemberCase value: lower_case - key: readability-identifier-naming.MemberPrefix value: '' - key: readability-identifier-naming.MemberSuffix value: '' - key: readability-identifier-naming.NamespaceCase value: lower_case - key: readability-identifier-naming.NamespacePrefix value: '' - key: readability-identifier-naming.NamespaceSuffix value: '' - key: readability-identifier-naming.PrivateMemberCase value: lower_case - key: readability-identifier-naming.PrivateMemberPrefix value: '' - key: readability-identifier-naming.PrivateMemberSuffix value: '_' - key: readability-identifier-naming.ProtectedMemberCase value: lower_case - key: readability-identifier-naming.ProtectedMemberPrefix value: '' - key: readability-identifier-naming.ProtectedMemberSuffix value: '_' - key: readability-identifier-naming.StaticConstantCase value: CamelCase - key: readability-identifier-naming.StaticConstantPrefix value: 'k' - key: readability-identifier-naming.StaticConstantSuffix value: '' - key: readability-identifier-naming.StructCase value: lower_case - key: readability-identifier-naming.StructPrefix value: '' - key: readability-identifier-naming.StructSuffix value: '' - key: readability-identifier-naming.TypeAliasCase value: lower_case - key: readability-identifier-naming.TypeAliasPrefix value: '' - key: readability-identifier-naming.TypeAliasSuffix value: '' - key: readability-identifier-naming.TypeTemplateParameterCase value: CamelCase - key: readability-identifier-naming.TypeTemplateParameterPrefix value: '' - key: readability-identifier-naming.TypeTemplateParameterSuffix value: '' - key: readability-identifier-naming.TypedefCase value: lower_case - key: readability-identifier-naming.TypedefPrefix value: '' - key: readability-identifier-naming.TypedefSuffix value: '' - key: readability-identifier-naming.VariableCase value: lower_case - key: readability-identifier-naming.VariablePrefix value: '' - key: readability-identifier-naming.VariableSuffix value: '' ...
0
rapidsai_public_repos/raft
rapidsai_public_repos/raft/cpp/.clang-format
--- # Refer to the following link for the explanation of each params: # http://releases.llvm.org/8.0.0/tools/clang/docs/ClangFormatStyleOptions.html Language: Cpp # BasedOnStyle: Google AccessModifierOffset: -1 AlignAfterOpenBracket: Align AlignConsecutiveAssignments: true AlignConsecutiveBitFields: true AlignConsecutiveDeclarations: false AlignConsecutiveMacros: true AlignEscapedNewlines: Left AlignOperands: true AlignTrailingComments: true AllowAllArgumentsOnNextLine: true AllowAllConstructorInitializersOnNextLine: true AllowAllParametersOfDeclarationOnNextLine: true AllowShortBlocksOnASingleLine: true AllowShortCaseLabelsOnASingleLine: true AllowShortEnumsOnASingleLine: true AllowShortFunctionsOnASingleLine: All AllowShortIfStatementsOnASingleLine: true AllowShortLambdasOnASingleLine: true AllowShortLoopsOnASingleLine: false # This is deprecated AlwaysBreakAfterDefinitionReturnType: None AlwaysBreakAfterReturnType: None AlwaysBreakBeforeMultilineStrings: true AlwaysBreakTemplateDeclarations: Yes BinPackArguments: false BinPackParameters: false BraceWrapping: AfterClass: false AfterControlStatement: false AfterEnum: false AfterFunction: false AfterNamespace: false AfterObjCDeclaration: false AfterStruct: false AfterUnion: false AfterExternBlock: false BeforeCatch: false BeforeElse: false IndentBraces: false # disabling the below splits, else, they'll just add to the vertical length of source files! SplitEmptyFunction: false SplitEmptyRecord: false SplitEmptyNamespace: false BreakAfterJavaFieldAnnotations: false BreakBeforeBinaryOperators: None BreakBeforeBraces: WebKit BreakBeforeInheritanceComma: false BreakBeforeTernaryOperators: true BreakConstructorInitializersBeforeComma: false BreakConstructorInitializers: BeforeColon BreakInheritanceList: BeforeColon BreakStringLiterals: true ColumnLimit: 100 CommentPragmas: '^ IWYU pragma:' CompactNamespaces: false ConstructorInitializerAllOnOneLineOrOnePerLine: true # Kept the below 2 to be the same as `IndentWidth` to keep everything uniform ConstructorInitializerIndentWidth: 2 ContinuationIndentWidth: 2 Cpp11BracedListStyle: true DerivePointerAlignment: false DisableFormat: false ExperimentalAutoDetectBinPacking: false FixNamespaceComments: true ForEachMacros: - foreach - Q_FOREACH - BOOST_FOREACH IncludeBlocks: Preserve IncludeIsMainRegex: '([-_](test|unittest))?$' IndentCaseLabels: true IndentPPDirectives: None IndentWidth: 2 IndentWrappedFunctionNames: false JavaScriptQuotes: Leave JavaScriptWrapImports: true KeepEmptyLinesAtTheStartOfBlocks: false MacroBlockBegin: '' MacroBlockEnd: '' MaxEmptyLinesToKeep: 1 NamespaceIndentation: None ObjCBinPackProtocolList: Never ObjCBlockIndentWidth: 2 ObjCSpaceAfterProperty: false ObjCSpaceBeforeProtocolList: true PenaltyBreakAssignment: 2 PenaltyBreakBeforeFirstCallParameter: 1 PenaltyBreakComment: 300 PenaltyBreakFirstLessLess: 120 PenaltyBreakString: 1000 PenaltyBreakTemplateDeclaration: 10 PenaltyExcessCharacter: 1000000 PenaltyReturnTypeOnItsOwnLine: 200 PointerAlignment: Left RawStringFormats: - Language: Cpp Delimiters: - cc - CC - cpp - Cpp - CPP - 'c++' - 'C++' CanonicalDelimiter: '' - Language: TextProto Delimiters: - pb - PB - proto - PROTO EnclosingFunctions: - EqualsProto - EquivToProto - PARSE_PARTIAL_TEXT_PROTO - PARSE_TEST_PROTO - PARSE_TEXT_PROTO - ParseTextOrDie - ParseTextProtoOrDie CanonicalDelimiter: '' BasedOnStyle: google # Enabling comment reflow causes doxygen comments to be messed up in their formats! ReflowComments: true SortIncludes: true SortUsingDeclarations: true SpaceAfterCStyleCast: false SpaceAfterTemplateKeyword: true SpaceBeforeAssignmentOperators: true SpaceBeforeCpp11BracedList: false SpaceBeforeCtorInitializerColon: true SpaceBeforeInheritanceColon: true SpaceBeforeParens: ControlStatements SpaceBeforeRangeBasedForLoopColon: true SpaceBeforeSquareBrackets: false SpaceInEmptyBlock: false SpaceInEmptyParentheses: false SpacesBeforeTrailingComments: 2 SpacesInAngles: false SpacesInConditionalStatement: false SpacesInContainerLiterals: true SpacesInCStyleCastParentheses: false SpacesInParentheses: false SpacesInSquareBrackets: false Standard: c++17 StatementMacros: - Q_UNUSED - QT_REQUIRE_VERSION # Be consistent with indent-width, even for people who use tab for indentation! TabWidth: 2 UseTab: Never
0
rapidsai_public_repos/raft/cpp
rapidsai_public_repos/raft/cpp/include/raft.hpp
/* * Copyright (c) 2020-2022, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ /** * This file is deprecated and will be removed in a future release. */ #include "raft/core/device_mdarray.hpp" #include "raft/core/device_mdspan.hpp" #include "raft/core/device_span.hpp" #include "raft/core/handle.hpp" #include <string> namespace raft { /* Function for testing RAFT include * * @return message indicating RAFT has been included successfully*/ inline std::string test_raft() { std::string status = "RAFT Setup successfully"; return status; } } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft_runtime
rapidsai_public_repos/raft/cpp/include/raft_runtime/random/rmat_rectangular_generator.hpp
/* * Copyright (c) 2022-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #include <cstdint> #include <raft/core/resources.hpp> #include <raft/random/rng_state.hpp> namespace raft::runtime::random { /** * @defgroup rmat_runtime RMAT Runtime API * @{ */ #define FUNC_DECL(IdxT, ProbT) \ void rmat_rectangular_gen(raft::resources const& handle, \ IdxT* out, \ IdxT* out_src, \ IdxT* out_dst, \ const ProbT* theta, \ IdxT r_scale, \ IdxT c_scale, \ IdxT n_edges, \ raft::random::RngState& r) FUNC_DECL(int, float); FUNC_DECL(int64_t, float); FUNC_DECL(int, double); FUNC_DECL(int64_t, double); #undef FUNC_DECL /** @} */ // end group rmat_runtime } // namespace raft::runtime::random
0
rapidsai_public_repos/raft/cpp/include/raft_runtime
rapidsai_public_repos/raft/cpp/include/raft_runtime/distance/pairwise_distance.hpp
/* * Copyright (c) 2022-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #include <raft/distance/distance_types.hpp> namespace raft::runtime::distance { /** * @defgroup pairwise_distance_runtime Pairwise Distances Runtime API * @{ */ void pairwise_distance(raft::resources const& handle, float* x, float* y, float* dists, int m, int n, int k, raft::distance::DistanceType metric, bool isRowMajor, float metric_arg); void pairwise_distance(raft::resources const& handle, double* x, double* y, double* dists, int m, int n, int k, raft::distance::DistanceType metric, bool isRowMajor, float metric_arg); /** @} */ // end group pairwise_distance_runtime } // namespace raft::runtime::distance
0
rapidsai_public_repos/raft/cpp/include/raft_runtime
rapidsai_public_repos/raft/cpp/include/raft_runtime/distance/fused_l2_nn.hpp
/* * Copyright (c) 2022-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #include <raft/core/resources.hpp> #include <raft/distance/distance_types.hpp> namespace raft::runtime::distance { /** * @defgroup fused_l2_nn_min_arg_runtime Fused L2 1NN Runtime API * @{ */ /** * @brief Wrapper around fusedL2NN with minimum reduction operators. * * fusedL2NN cannot be compiled in the distance library due to the lambda * operators, so this wrapper covers the most common case (minimum). * * @param[in] handle raft handle * @param[out] min will contain the reduced output (Length = `m`) * (on device) * @param[in] x first matrix. Row major. Dim = `m x k`. * (on device). * @param[in] y second matrix. Row major. Dim = `n x k`. * (on device). * @param[in] m gemm m * @param[in] n gemm n * @param[in] k gemm k * @param[in] sqrt Whether the output `minDist` should contain L2-sqrt */ void fused_l2_nn_min_arg(raft::resources const& handle, int* min, const float* x, const float* y, int m, int n, int k, bool sqrt); void fused_l2_nn_min_arg(raft::resources const& handle, int* min, const double* x, const double* y, int m, int n, int k, bool sqrt); /** @} */ // end group fused_l2_nn_min_arg_runtime } // end namespace raft::runtime::distance
0
rapidsai_public_repos/raft/cpp/include/raft_runtime
rapidsai_public_repos/raft/cpp/include/raft_runtime/neighbors/ivf_pq.hpp
/* * Copyright (c) 2022-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/neighbors/ivf_pq_types.hpp> namespace raft::runtime::neighbors::ivf_pq { // We define overloads for build and extend with void return type. This is used in the Cython // wrappers, where exception handling is not compatible with return type that has nontrivial // constructor. #define RAFT_DECL_BUILD_EXTEND(T, IdxT) \ [[nodiscard]] raft::neighbors::ivf_pq::index<IdxT> build( \ raft::resources const& handle, \ const raft::neighbors::ivf_pq::index_params& params, \ raft::device_matrix_view<const T, IdxT, row_major> dataset); \ \ void build(raft::resources const& handle, \ const raft::neighbors::ivf_pq::index_params& params, \ raft::device_matrix_view<const T, IdxT, row_major> dataset, \ raft::neighbors::ivf_pq::index<IdxT>* idx); \ \ [[nodiscard]] raft::neighbors::ivf_pq::index<IdxT> extend( \ raft::resources const& handle, \ raft::device_matrix_view<const T, IdxT, row_major> new_vectors, \ std::optional<raft::device_vector_view<const IdxT, IdxT>> new_indices, \ const raft::neighbors::ivf_pq::index<IdxT>& idx); \ \ void extend(raft::resources const& handle, \ raft::device_matrix_view<const T, IdxT, row_major> new_vectors, \ std::optional<raft::device_vector_view<const IdxT, IdxT>> new_indices, \ raft::neighbors::ivf_pq::index<IdxT>* idx); RAFT_DECL_BUILD_EXTEND(float, int64_t); RAFT_DECL_BUILD_EXTEND(int8_t, int64_t); RAFT_DECL_BUILD_EXTEND(uint8_t, int64_t); #undef RAFT_DECL_BUILD_EXTEND #define RAFT_DECL_SEARCH(T, IdxT) \ void search(raft::resources const& handle, \ const raft::neighbors::ivf_pq::search_params& params, \ const raft::neighbors::ivf_pq::index<IdxT>& idx, \ raft::device_matrix_view<const T, IdxT, row_major> queries, \ raft::device_matrix_view<IdxT, IdxT, row_major> neighbors, \ raft::device_matrix_view<float, IdxT, row_major> distances); RAFT_DECL_SEARCH(float, int64_t); RAFT_DECL_SEARCH(int8_t, int64_t); RAFT_DECL_SEARCH(uint8_t, int64_t); #undef RAFT_DECL_SEARCH /** * Save the index to file. * * Experimental, both the API and the serialization format are subject to change. * * @param[in] handle the raft handle * @param[in] filename the filename for saving the index * @param[in] index IVF-PQ index * */ void serialize(raft::resources const& handle, const std::string& filename, const raft::neighbors::ivf_pq::index<int64_t>& index); /** * Load index from file. * * Experimental, both the API and the serialization format are subject to change. * * @param[in] handle the raft handle * @param[in] filename the name of the file that stores the index * @param[in] index IVF-PQ index * */ void deserialize(raft::resources const& handle, const std::string& filename, raft::neighbors::ivf_pq::index<int64_t>* index); } // namespace raft::runtime::neighbors::ivf_pq
0
rapidsai_public_repos/raft/cpp/include/raft_runtime
rapidsai_public_repos/raft/cpp/include/raft_runtime/neighbors/ivf_flat.hpp
/* * Copyright (c) 2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/neighbors/ivf_flat_types.hpp> #include <string> namespace raft::runtime::neighbors::ivf_flat { // We define overloads for build and extend with void return type. This is used in the Cython // wrappers, where exception handling is not compatible with return type that has nontrivial // constructor. #define RAFT_INST_BUILD_EXTEND(T, IdxT) \ auto build(raft::resources const& handle, \ const raft::neighbors::ivf_flat::index_params& params, \ raft::device_matrix_view<const T, IdxT, row_major> dataset) \ ->raft::neighbors::ivf_flat::index<T, IdxT>; \ \ auto extend(raft::resources const& handle, \ raft::device_matrix_view<const T, IdxT, row_major> new_vectors, \ std::optional<raft::device_vector_view<const IdxT, IdxT>> new_indices, \ const raft::neighbors::ivf_flat::index<T, IdxT>& orig_index) \ ->raft::neighbors::ivf_flat::index<T, IdxT>; \ \ void build(raft::resources const& handle, \ const raft::neighbors::ivf_flat::index_params& params, \ raft::device_matrix_view<const T, IdxT, row_major> dataset, \ raft::neighbors::ivf_flat::index<T, IdxT>& idx); \ \ void extend(raft::resources const& handle, \ raft::device_matrix_view<const T, IdxT, row_major> new_vectors, \ std::optional<raft::device_vector_view<const IdxT, IdxT>> new_indices, \ raft::neighbors::ivf_flat::index<T, IdxT>* idx); \ \ void serialize_file(raft::resources const& handle, \ const std::string& filename, \ const raft::neighbors::ivf_flat::index<T, IdxT>& index); \ \ void deserialize_file(raft::resources const& handle, \ const std::string& filename, \ raft::neighbors::ivf_flat::index<T, IdxT>* index); \ void serialize(raft::resources const& handle, \ std::string& str, \ const raft::neighbors::ivf_flat::index<T, IdxT>& index); \ void deserialize(raft::resources const& handle, \ const std::string& str, \ raft::neighbors::ivf_flat::index<T, IdxT>*); RAFT_INST_BUILD_EXTEND(float, int64_t) RAFT_INST_BUILD_EXTEND(int8_t, int64_t) RAFT_INST_BUILD_EXTEND(uint8_t, int64_t) #undef RAFT_INST_BUILD_EXTEND #define RAFT_INST_SEARCH(T, IdxT) \ void search(raft::resources const&, \ raft::neighbors::ivf_flat::search_params const&, \ raft::neighbors::ivf_flat::index<T, IdxT> const&, \ raft::device_matrix_view<const T, IdxT, row_major>, \ raft::device_matrix_view<IdxT, IdxT, row_major>, \ raft::device_matrix_view<float, IdxT, row_major>); RAFT_INST_SEARCH(float, int64_t); RAFT_INST_SEARCH(int8_t, int64_t); RAFT_INST_SEARCH(uint8_t, int64_t); #undef RAFT_INST_SEARCH } // namespace raft::runtime::neighbors::ivf_flat
0
rapidsai_public_repos/raft/cpp/include/raft_runtime
rapidsai_public_repos/raft/cpp/include/raft_runtime/neighbors/refine.hpp
/* * Copyright (c) 2022-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/core/device_mdspan.hpp> #include <raft/core/resources.hpp> // #include <raft/core/host_mdspan.hpp> namespace raft::runtime::neighbors { #define RAFT_INST_REFINE(IDX_T, DATA_T) \ void refine(raft::resources const& handle, \ raft::device_matrix_view<const DATA_T, int64_t, row_major> dataset, \ raft::device_matrix_view<const DATA_T, int64_t, row_major> queries, \ raft::device_matrix_view<const IDX_T, int64_t, row_major> neighbor_candidates, \ raft::device_matrix_view<IDX_T, int64_t, row_major> indices, \ raft::device_matrix_view<float, int64_t, row_major> distances, \ distance::DistanceType metric); \ \ void refine(raft::resources const& handle, \ raft::host_matrix_view<const DATA_T, int64_t, row_major> dataset, \ raft::host_matrix_view<const DATA_T, int64_t, row_major> queries, \ raft::host_matrix_view<const IDX_T, int64_t, row_major> neighbor_candidates, \ raft::host_matrix_view<IDX_T, int64_t, row_major> indices, \ raft::host_matrix_view<float, int64_t, row_major> distances, \ distance::DistanceType metric); RAFT_INST_REFINE(int64_t, float); RAFT_INST_REFINE(int64_t, uint8_t); RAFT_INST_REFINE(int64_t, int8_t); #undef RAFT_INST_REFINE } // namespace raft::runtime::neighbors
0
rapidsai_public_repos/raft/cpp/include/raft_runtime
rapidsai_public_repos/raft/cpp/include/raft_runtime/neighbors/brute_force.hpp
/* * Copyright (c) 2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/core/device_mdspan.hpp> #include <raft/core/resources.hpp> namespace raft::runtime::neighbors::brute_force { #define RAFT_INST_BFKNN(IDX_T, DATA_T, MATRIX_IDX_T, INDEX_LAYOUT, SEARCH_LAYOUT) \ void knn(raft::resources const& handle, \ raft::device_matrix_view<const DATA_T, MATRIX_IDX_T, INDEX_LAYOUT> index, \ raft::device_matrix_view<const DATA_T, MATRIX_IDX_T, SEARCH_LAYOUT> search, \ raft::device_matrix_view<IDX_T, MATRIX_IDX_T, row_major> indices, \ raft::device_matrix_view<DATA_T, MATRIX_IDX_T, row_major> distances, \ distance::DistanceType metric = distance::DistanceType::L2Unexpanded, \ std::optional<float> metric_arg = std::make_optional<float>(2.0f), \ std::optional<IDX_T> global_id_offset = std::nullopt); RAFT_INST_BFKNN(int64_t, float, int64_t, raft::row_major, raft::row_major); #undef RAFT_INST_BFKNN } // namespace raft::runtime::neighbors::brute_force
0
rapidsai_public_repos/raft/cpp/include/raft_runtime
rapidsai_public_repos/raft/cpp/include/raft_runtime/neighbors/cagra.hpp
/* * Copyright (c) 2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/neighbors/cagra_types.hpp> #include <raft/neighbors/ivf_pq_types.hpp> #include <string> #include <raft/core/device_mdspan.hpp> #include <raft/core/host_device_accessor.hpp> #include <raft/core/mdspan.hpp> namespace raft::runtime::neighbors::cagra { // Using device and host_matrix_view avoids needing to typedef mutltiple mdspans based on accessors #define RAFT_INST_CAGRA_FUNCS(T, IdxT) \ auto build(raft::resources const& handle, \ const raft::neighbors::cagra::index_params& params, \ raft::device_matrix_view<const T, int64_t, row_major> dataset) \ ->raft::neighbors::cagra::index<T, IdxT>; \ \ auto build(raft::resources const& handle, \ const raft::neighbors::cagra::index_params& params, \ raft::host_matrix_view<const T, int64_t, row_major> dataset) \ ->raft::neighbors::cagra::index<T, IdxT>; \ \ void build_device(raft::resources const& handle, \ const raft::neighbors::cagra::index_params& params, \ raft::device_matrix_view<const T, int64_t, row_major> dataset, \ raft::neighbors::cagra::index<T, IdxT>& idx); \ \ void build_host(raft::resources const& handle, \ const raft::neighbors::cagra::index_params& params, \ raft::host_matrix_view<const T, int64_t, row_major> dataset, \ raft::neighbors::cagra::index<T, IdxT>& idx); \ \ void search(raft::resources const& handle, \ raft::neighbors::cagra::search_params const& params, \ const raft::neighbors::cagra::index<T, IdxT>& index, \ raft::device_matrix_view<const T, int64_t, row_major> queries, \ raft::device_matrix_view<IdxT, int64_t, row_major> neighbors, \ raft::device_matrix_view<float, int64_t, row_major> distances); \ void serialize_file(raft::resources const& handle, \ const std::string& filename, \ const raft::neighbors::cagra::index<T, IdxT>& index, \ bool include_dataset = true); \ \ void deserialize_file(raft::resources const& handle, \ const std::string& filename, \ raft::neighbors::cagra::index<T, IdxT>* index); \ void serialize(raft::resources const& handle, \ std::string& str, \ const raft::neighbors::cagra::index<T, IdxT>& index, \ bool include_dataset = true); \ \ void deserialize(raft::resources const& handle, \ const std::string& str, \ raft::neighbors::cagra::index<T, IdxT>* index); RAFT_INST_CAGRA_FUNCS(float, uint32_t); RAFT_INST_CAGRA_FUNCS(int8_t, uint32_t); RAFT_INST_CAGRA_FUNCS(uint8_t, uint32_t); #undef RAFT_INST_CAGRA_FUNCS #define RAFT_INST_CAGRA_OPTIMIZE(IdxT) \ void optimize_device(raft::resources const& res, \ raft::device_matrix_view<IdxT, int64_t, row_major> knn_graph, \ raft::host_matrix_view<IdxT, int64_t, row_major> new_graph); \ \ void optimize_host(raft::resources const& res, \ raft::host_matrix_view<IdxT, int64_t, row_major> knn_graph, \ raft::host_matrix_view<IdxT, int64_t, row_major> new_graph); RAFT_INST_CAGRA_OPTIMIZE(uint32_t); #undef RAFT_INST_CAGRA_OPTIMIZE } // namespace raft::runtime::neighbors::cagra
0
rapidsai_public_repos/raft/cpp/include/raft_runtime
rapidsai_public_repos/raft/cpp/include/raft_runtime/matrix/select_k.hpp
/* * Copyright (c) 2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/core/device_mdspan.hpp> #include <raft/core/resources.hpp> #include <optional> namespace raft::runtime::matrix { void select_k(const resources& handle, raft::device_matrix_view<const float, int64_t, row_major> in_val, std::optional<raft::device_matrix_view<const int64_t, int64_t, row_major>> in_idx, raft::device_matrix_view<float, int64_t, row_major> out_val, raft::device_matrix_view<int64_t, int64_t, row_major> out_idx, bool select_min); } // namespace raft::runtime::matrix
0
rapidsai_public_repos/raft/cpp/include/raft_runtime
rapidsai_public_repos/raft/cpp/include/raft_runtime/cluster/kmeans.hpp
/* * Copyright (c) 2022-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #include <raft/core/device_mdspan.hpp> #include <raft/core/host_mdspan.hpp> #include <raft/core/resources.hpp> #include <raft/distance/distance_types.hpp> #include <raft/cluster/kmeans_types.hpp> namespace raft::runtime::cluster::kmeans { /** * @defgroup kmeans_runtime Kmeans Runtime API * @{ */ void update_centroids(raft::resources const& handle, const float* X, int n_samples, int n_features, int n_clusters, const float* sample_weights, const float* centroids, const int* labels, float* new_centroids, float* weight_per_cluster); void update_centroids(raft::resources const& handle, const double* X, int n_samples, int n_features, int n_clusters, const double* sample_weights, const double* centroids, const int* labels, double* new_centroids, double* weight_per_cluster); void fit(raft::resources const& handle, const raft::cluster::kmeans::KMeansParams& params, raft::device_matrix_view<const float, int, row_major> X, std::optional<raft::device_vector_view<const float, int>> sample_weight, raft::device_matrix_view<float, int, row_major> centroids, raft::host_scalar_view<float, int> inertia, raft::host_scalar_view<int, int> n_iter); void fit(raft::resources const& handle, const raft::cluster::kmeans::KMeansParams& params, raft::device_matrix_view<const double, int, row_major> X, std::optional<raft::device_vector_view<const double, int>> sample_weight, raft::device_matrix_view<double, int, row_major> centroids, raft::host_scalar_view<double, int> inertia, raft::host_scalar_view<int, int> n_iter); void init_plus_plus(raft::resources const& handle, const raft::cluster::kmeans::KMeansParams& params, raft::device_matrix_view<const float, int, row_major> X, raft::device_matrix_view<float, int, row_major> centroids); void init_plus_plus(raft::resources const& handle, const raft::cluster::kmeans::KMeansParams& params, raft::device_matrix_view<const double, int, row_major> X, raft::device_matrix_view<double, int, row_major> centroids); void cluster_cost(raft::resources const& handle, const float* X, int n_samples, int n_features, int n_clusters, const float* centroids, float* cost); void cluster_cost(raft::resources const& handle, const double* X, int n_samples, int n_features, int n_clusters, const double* centroids, double* cost); /** @} */ // end group kmeans_runtime } // namespace raft::runtime::cluster::kmeans
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/seive.hpp
/* * Copyright (c) 2019-2022, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/util/cuda_utils.cuh> #include <vector> // Taken from: // https://github.com/teju85/programming/blob/master/euler/include/seive.h namespace raft { namespace common { /** * @brief Implementation of 'Seive of Eratosthenes' */ class Seive { public: /** * @param _num number of integers for which seive is needed */ Seive(unsigned _num) { N = _num; generateSeive(); } /** * @brief Check whether a number is prime or not * @param num number to be checked * @return true if the 'num' is prime, else false */ bool isPrime(unsigned num) const { unsigned mask, pos; if (num <= 1) { return false; } if (num == 2) { return true; } if (!(num & 1)) { return false; } getMaskPos(num, mask, pos); return (seive[pos] & mask); } private: void generateSeive() { auto sqN = fastIntSqrt(N); auto size = raft::ceildiv<unsigned>(N, sizeof(unsigned) * 8); seive.resize(size); // assume all to be primes initially for (auto& itr : seive) { itr = 0xffffffffu; } unsigned cid = 0; unsigned cnum = getNum(cid); while (cnum <= sqN) { do { ++cid; cnum = getNum(cid); if (isPrime(cnum)) { break; } } while (cnum <= sqN); auto cnum2 = cnum << 1; // 'unmark' all the 'odd' multiples of the current prime for (unsigned i = 3, num = i * cnum; num <= N; i += 2, num += cnum2) { unmark(num); } } } unsigned getId(unsigned num) const { return (num >> 1); } unsigned getNum(unsigned id) const { if (id == 0) { return 2; } return ((id << 1) + 1); } void getMaskPos(unsigned num, unsigned& mask, unsigned& pos) const { pos = getId(num); mask = 1 << (pos & 0x1f); pos >>= 5; } void unmark(unsigned num) { unsigned mask, pos; getMaskPos(num, mask, pos); seive[pos] &= ~mask; } // REF: http://www.azillionmonkeys.com/qed/ulerysqroot.pdf unsigned fastIntSqrt(unsigned val) { unsigned g = 0; auto bshft = 15u, b = 1u << bshft; do { unsigned temp = ((g << 1) + b) << bshft--; if (val >= temp) { g += b; val -= temp; } } while (b >>= 1); return g; } /** find all primes till this number */ unsigned N; /** the seive */ std::vector<unsigned> seive; }; }; // namespace common }; // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/fast_int_div.cuh
/* * Copyright (c) 2020-2022, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/util/cuda_utils.cuh> #include <stdint.h> namespace raft::util { /** * @brief Perform fast integer division and modulo using a known divisor * From Hacker's Delight, Second Edition, Chapter 10 * * @note This currently only supports 32b signed integers * @todo Extend support for signed divisors */ struct FastIntDiv { /** * @defgroup HostMethods Ctor's that are accessible only from host * @{ * @brief Host-only ctor's * @param _d the divisor */ FastIntDiv(int _d) : d(_d) { computeScalars(); } FastIntDiv& operator=(int _d) { d = _d; computeScalars(); return *this; } /** @} */ /** * @defgroup DeviceMethods Ctor's which even the device-side can access * @{ * @brief host and device ctor's * @param other source object to be copied from */ HDI FastIntDiv(const FastIntDiv& other) : d(other.d), m(other.m), p(other.p) {} HDI FastIntDiv& operator=(const FastIntDiv& other) { d = other.d; m = other.m; p = other.p; return *this; } /** @} */ /** divisor */ int d; /** the term 'm' as found in the reference chapter */ unsigned m; /** the term 'p' as found in the reference chapter */ int p; private: void computeScalars() { if (d == 1) { m = 0; p = 1; return; } else if (d < 0) { ASSERT(false, "FastIntDiv: division by negative numbers not supported!"); } else if (d == 0) { ASSERT(false, "FastIntDiv: got division by zero!"); } int64_t nc = ((1LL << 31) / d) * d - 1; p = 31; int64_t twoP, rhs; do { ++p; twoP = 1LL << p; rhs = nc * (d - twoP % d); } while (twoP <= rhs); m = (twoP + d - twoP % d) / d; } }; // struct FastIntDiv /** * @brief Division overload, so that FastIntDiv can be transparently switched * to even on device * @param n numerator * @param divisor the denominator * @return the quotient */ HDI int operator/(int n, const FastIntDiv& divisor) { if (divisor.d == 1) return n; int ret = (int64_t(divisor.m) * int64_t(n)) >> divisor.p; if (n < 0) ++ret; return ret; } /** * @brief Modulo overload, so that FastIntDiv can be transparently switched * to even on device * @param n numerator * @param divisor the denominator * @return the remainder */ HDI int operator%(int n, const FastIntDiv& divisor) { int quotient = n / divisor; int remainder = n - quotient * divisor.d; return remainder; } }; // namespace raft::util
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/arch.cuh
/* * Copyright (c) 2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/util/cuda_rt_essentials.hpp> // RAFT_CUDA_TRY namespace raft::util::arch { /* raft::util::arch provides the following facilities: * * - raft::util::arch::SM_XX : hardcoded compile-time constants for various compute * architectures. The values raft::util::arch::SM_min and raft::util::arch::SM_future * represent architectures that are always smaller and larger (respectively) * than any architecture that can be encountered in practice. * * - raft::util::arch::SM_compute_arch : a compile-time value for the *current* * compute architecture that a kernel is compiled with. It can only be used * inside kernels with a template argument. * * - raft::util::arch::kernel_virtual_arch : a function that computes at *run-time* * which version of a kernel will launch (i.e., it will return the virtual compute * architecture of the version of the kernel that it was compiled for which * will be launched by the driver). * * - raft::util::arch::SM_range : a compile-time value to represent an open interval * of compute architectures. This can be used to check if the current * compile-time architecture is in a specified compatibility range. */ // detail::SM_generic is a template to create a generic compile-time SM // architecture constant. namespace detail { template <int n> struct SM_generic { public: __host__ __device__ constexpr int value() const { return n; } }; } // namespace detail // A list of architectures that RAPIDS explicitly builds for (SM60, ..., SM90) // and SM_MIN and SM_FUTURE, that allow specifying an open interval of // compatible compute architectures. using SM_min = detail::SM_generic<350>; using SM_60 = detail::SM_generic<600>; using SM_70 = detail::SM_generic<700>; using SM_75 = detail::SM_generic<750>; using SM_80 = detail::SM_generic<800>; using SM_86 = detail::SM_generic<860>; using SM_90 = detail::SM_generic<900>; using SM_future = detail::SM_generic<99999>; // This is a type that uses the __CUDA_ARCH__ macro to obtain the compile-time // compute architecture. It can only be used where __CUDA_ARCH__ is defined, // i.e., inside a __global__ function template. struct SM_compute_arch { template <int dummy = 0> __device__ constexpr int value() const { #ifdef __CUDA_ARCH__ return __CUDA_ARCH__; #else // This function should not be called in host code (because __CUDA_ARCH__ is // not defined). This function is constexpr and thus can be called in host // code (due to the --expt-relaxed-constexpr compile flag). We would like to // provide an intelligible error message when this function is called in // host code, which we do below. // // To make sure the static_assert only fires in host code, we use a dummy // template parameter as described in P2593: // https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2022/p2593r0.html static_assert(dummy != 0, "SM_compute_arch.value() is only callable from a __global__ function template. " "A way to create a function template is by adding 'template <int dummy = 0>'."); return -1; #endif } }; // A runtime value for the actual compute architecture of a kernel. // // A single kernel can be compiled for several "virtual" compute architectures. // When a program runs, the driver picks the version of the kernel that most // closely matches the current hardware. This struct reflects the virtual // compute architecture of the version of the kernel that the driver picks when // the kernel runs. struct SM_runtime { friend SM_runtime kernel_virtual_arch(void*); private: const int _version; SM_runtime(int version) : _version(version) {} public: __host__ __device__ int value() const { return _version; } }; // Computes which virtual compute architecture the given kernel was compiled for, // driver picks the version of the kernel that closely matches the current hardware. // // Semantics are described above in the documentation of SM_runtime. // // This function requires a pointer to the kernel that will run. Other methods // to determine the architecture (that do not require a pointer) can be error // prone. See: // https://github.com/NVIDIA/cub/issues/545 inline SM_runtime kernel_virtual_arch(void* kernel) { cudaFuncAttributes attributes; RAFT_CUDA_TRY(cudaFuncGetAttributes(&attributes, kernel)); return SM_runtime(10 * attributes.ptxVersion); } // SM_range represents a range of SM architectures. It can be used to // conditionally compile a kernel. template <typename SM_MIN, typename SM_MAX> struct SM_range { private: const SM_MIN _min; const SM_MAX _max; public: __host__ __device__ constexpr SM_range(SM_MIN min, SM_MAX max) : _min(min), _max(max) {} __host__ __device__ constexpr SM_range() : _min(SM_MIN()), _max(SM_MAX()) {} template <typename SM_t> __host__ __device__ constexpr bool contains(SM_t current) const { return _min.value() <= current.value() && current.value() < _max.value(); } }; } // namespace raft::util::arch
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/memory_pool-ext.hpp
/* * Copyright (c) 2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <cstddef> // size_t #include <memory> // std::unique_ptr #include <rmm/mr/device/device_memory_resource.hpp> // rmm::mr::device_memory_resource namespace raft { std::unique_ptr<rmm::mr::device_memory_resource> get_pool_memory_resource( rmm::mr::device_memory_resource*& mr, size_t initial_size); } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/integer_utils.hpp
/* * Copyright 2019 BlazingDB, Inc. * Copyright 2019 Eyal Rozenberg <eyalroz@blazingdb.com> * Copyright (c) 2020-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once /** * Utility code involving integer arithmetic * */ #include <limits> #include <raft/core/detail/macros.hpp> #include <stdexcept> #include <type_traits> namespace raft { //! Utility functions /** * Finds the smallest integer not less than `number_to_round` and modulo `S` is * zero. This function assumes that `number_to_round` is non-negative and * `modulus` is positive. */ template <typename S> constexpr inline S round_up_safe(S number_to_round, S modulus) { auto remainder = number_to_round % modulus; if (remainder == 0) { return number_to_round; } auto rounded_up = number_to_round - remainder + modulus; if (rounded_up < number_to_round) { throw std::invalid_argument("Attempt to round up beyond the type's maximum value"); } return rounded_up; } /** * Finds the largest integer not greater than `number_to_round` and modulo `S` is * zero. This function assumes that `number_to_round` is non-negative and * `modulus` is positive. */ template <typename S> inline S round_down_safe(S number_to_round, S modulus) { auto remainder = number_to_round % modulus; auto rounded_down = number_to_round - remainder; return rounded_down; } /** * Divides the left-hand-side by the right-hand-side, rounding up * to an integral multiple of the right-hand-side, e.g. (9,5) -> 2 , (10,5) -> 2, (11,5) -> 3. * * @param dividend the number to divide * @param divisor the number by which to divide * @return The least integer multiple of divisor which is greater than or equal to * the non-integral division dividend/divisor. * * @note sensitive to overflow, i.e. if dividend > std::numeric_limits<S>::max() - divisor, * the result will be incorrect */ template <typename S, typename T> constexpr inline S div_rounding_up_unsafe(const S& dividend, const T& divisor) noexcept { return (dividend + divisor - 1) / divisor; } namespace detail { template <typename I> constexpr inline I div_rounding_up_safe(std::integral_constant<bool, false>, I dividend, I divisor) noexcept { // TODO: This could probably be implemented faster return (dividend > divisor) ? 1 + div_rounding_up_unsafe(dividend - divisor, divisor) : (dividend > 0); } template <typename I> constexpr inline I div_rounding_up_safe(std::integral_constant<bool, true>, I dividend, I divisor) noexcept { auto quotient = dividend / divisor; auto remainder = dividend % divisor; return quotient + (remainder != 0); } } // namespace detail /** * Divides the left-hand-side by the right-hand-side, rounding up * to an integral multiple of the right-hand-side, e.g. (9,5) -> 2 , (10,5) -> 2, (11,5) -> 3. * * @param dividend the number to divide * @param divisor the number of by which to divide * @return The least integer multiple of divisor which is greater than or equal to * the non-integral division dividend/divisor. * * @note will not overflow, and may _or may not_ be slower than the intuitive * approach of using (dividend + divisor - 1) / divisor */ template <typename I> constexpr inline auto div_rounding_up_safe(I dividend, I divisor) noexcept -> std::enable_if_t<std::is_integral<I>::value, I> { using i_is_a_signed_type = std::integral_constant<bool, std::is_signed<I>::value>; return detail::div_rounding_up_safe(i_is_a_signed_type{}, dividend, divisor); } template <typename I> constexpr inline auto is_a_power_of_two(I val) noexcept -> std::enable_if_t<std::is_integral<I>::value, bool> { return (val != 0) && (((val - 1) & val) == 0); } /** * Given an integer `x`, return such `y` that `x <= y` and `is_a_power_of_two(y)`. * If such `y` does not exist in `T`, return zero. */ template <typename T> constexpr inline auto bound_by_power_of_two(T x) noexcept -> std::enable_if_t<std::is_integral<T>::value, T> { if (is_a_power_of_two(x)) { return x; } constexpr T kMaxUnsafe = std::numeric_limits<T>::max(); constexpr T kMaxSafe = is_a_power_of_two(kMaxUnsafe) ? kMaxUnsafe : (kMaxUnsafe >> 1); const T limited = std::min(x, kMaxSafe); T bound = T{1}; while (bound < limited) { bound <<= 1; } return bound < x ? T{0} : bound; } /** * @brief Return the absolute value of a number. * * This calls `std::abs()` which performs equivalent: `(value < 0) ? -value : value`. * * This was created to prevent compile errors calling `std::abs()` with unsigned integers. * An example compile error appears as follows: * @code{.pseudo} * error: more than one instance of overloaded function "std::abs" matches the argument list: * function "abs(int)" * function "std::abs(long)" * function "std::abs(long long)" * function "std::abs(double)" * function "std::abs(float)" * function "std::abs(long double)" * argument types are: (uint64_t) * @endcode * * Not all cases could be if-ed out using std::is_signed<T>::value and satisfy the compiler. * * @param val Numeric value can be either integer or float type. * @return Absolute value if value type is signed. */ template <typename T> constexpr inline auto absolute_value(T val) -> std::enable_if_t<std::is_signed<T>::value, T> { return std::abs(val); } // Unsigned type just returns itself. template <typename T> constexpr inline auto absolute_value(T val) -> std::enable_if_t<!std::is_signed<T>::value, T> { return val; } /** * @defgroup Check whether the numeric conversion is narrowing * * @tparam From source type * @tparam To destination type * @{ */ template <typename From, typename To, typename = void> struct is_narrowing : std::true_type {}; template <typename From, typename To> struct is_narrowing<From, To, std::void_t<decltype(To{std::declval<From>()})>> : std::false_type {}; /** @} */ /** Check whether the numeric conversion is narrowing */ template <typename From, typename To> inline constexpr bool is_narrowing_v = is_narrowing<From, To>::value; // NOLINT /** Wide multiplication of two unsigned 64-bit integers */ _RAFT_HOST_DEVICE inline void wmul_64bit(uint64_t& res_hi, uint64_t& res_lo, uint64_t a, uint64_t b) { #ifdef __CUDA_ARCH__ asm("mul.hi.u64 %0, %1, %2;" : "=l"(res_hi) : "l"(a), "l"(b)); asm("mul.lo.u64 %0, %1, %2;" : "=l"(res_lo) : "l"(a), "l"(b)); #else uint32_t a_hi, a_lo, b_hi, b_lo; a_hi = uint32_t(a >> 32); a_lo = uint32_t(a & uint64_t(0x00000000FFFFFFFF)); b_hi = uint32_t(b >> 32); b_lo = uint32_t(b & uint64_t(0x00000000FFFFFFFF)); uint64_t t0 = uint64_t(a_lo) * uint64_t(b_lo); uint64_t t1 = uint64_t(a_hi) * uint64_t(b_lo); uint64_t t2 = uint64_t(a_lo) * uint64_t(b_hi); uint64_t t3 = uint64_t(a_hi) * uint64_t(b_hi); uint64_t carry = 0, trial = 0; res_lo = t0; trial = res_lo + (t1 << 32); if (trial < res_lo) carry++; res_lo = trial; trial = res_lo + (t2 << 32); if (trial < res_lo) carry++; res_lo = trial; // No need to worry about carry in this addition res_hi = (t1 >> 32) + (t2 >> 32) + t3 + carry; #endif } } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/cache_util.cuh
/* * Copyright (c) 2019-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <cub/cub.cuh> #include <raft/util/cuda_utils.cuh> namespace raft { namespace cache { /** * @brief Collect vectors of data from the cache into a contiguous memory buffer. * * We assume contiguous memory layout for the output buffer, i.e. we get * column vectors into a column major out buffer, or row vectors into a row * major output buffer. * * On exit, the output array is filled the following way: * out[i + n_vec*k] = cache[i + n_vec * cache_idx[k]]), where i=0..n_vec-1, and * k = 0..n-1 where cache_idx[k] >= 0 * * We ignore vectors where cache_idx[k] < 0. * * @param [in] cache stores the cached data, size [n_vec x n_cached_vectors] * @param [in] n_vec number of elements in a cached vector * @param [in] cache_idx cache indices, size [n] * @param [in] n the number of elements that need to be collected * @param [out] out vectors collected from the cache, size [n_vec * n] */ template <typename math_t, typename idx_t, typename int_t> RAFT_KERNEL get_vecs(const math_t* cache, int_t n_vec, const idx_t* cache_idx, int_t n, math_t* out) { int tid = threadIdx.x + blockIdx.x * blockDim.x; int row = tid % n_vec; // row idx if (tid < n_vec * n) { size_t out_col = tid / n_vec; // col idx size_t cache_col = cache_idx[out_col]; if (!std::is_signed<idx_t>::value || cache_idx[out_col] >= 0) { if (row + out_col * n_vec < (size_t)n_vec * n) { out[tid] = cache[row + cache_col * n_vec]; } } } } /** * @brief Store vectors of data into the cache. * * Elements within a vector should be contiguous in memory (i.e. column vectors * for column major data storage, or row vectors of row major data). * * If tile_idx==nullptr then the operation is the opposite of get_vecs, * i.e. we store * cache[i + cache_idx[k]*n_vec] = tile[i + k*n_vec], for i=0..n_vec-1, k=0..n-1 * * If tile_idx != nullptr, then we permute the vectors from tile according * to tile_idx. This allows to store vectors from a buffer where the individual * vectors are not stored contiguously (but the elements of each vector shall * be contiguous): * cache[i + cache_idx[k]*n_vec] = tile[i + tile_idx[k]*n_vec], * for i=0..n_vec-1, k=0..n-1 * * @param [in] tile stores the data to be cashed cached, size [n_vec x n_tile] * @param [in] n_tile number of vectors in the input tile * @param [in] n_vec number of elements in a cached vector * @param [in] tile_idx indices of vectors that we want to store * @param [in] n number of vectos that we want to store (n <= n_tile) * @param [in] cache_idx cache indices, size [n], negative values are ignored * @param [inout] cache updated cache * @param [in] n_cache_vecs */ template <typename math_t> RAFT_KERNEL store_vecs(const math_t* tile, int n_tile, int n_vec, const int* tile_idx, int n, const int* cache_idx, math_t* cache, int n_cache_vecs) { int tid = threadIdx.x + blockIdx.x * blockDim.x; int row = tid % n_vec; // row idx if (tid < n_vec * n) { int tile_col = tid / n_vec; // col idx int data_col = tile_idx ? tile_idx[tile_col] : tile_col; int cache_col = cache_idx[tile_col]; // We ignore negative values. The rest of the checks should be fulfilled // if the cache is used properly if (cache_col >= 0 && cache_col < n_cache_vecs && data_col < n_tile) { cache[row + (size_t)cache_col * n_vec] = tile[row + (size_t)data_col * n_vec]; } } } /** * @brief Map a key to a cache set. * * @param key key to be hashed * @param n_cache_sets number of cache sets * @return index of the cache set [0..n_cache_set) */ int DI hash(int key, int n_cache_sets) { return key % n_cache_sets; } /** * @brief Binary search to find the first element in the array which is greater * equal than a given value. * @param [in] array sorted array of n numbers * @param [in] n length of the array * @param [in] val the value to search for * @return the index of the first element in the array for which * array[idx] >= value. If there is no such value, then return n. */ int DI arg_first_ge(const int* array, int n, int val) { int start = 0; int end = n - 1; if (array[0] == val) return 0; if (array[end] < val) return n; while (start + 1 < end) { int q = (start + end + 1) / 2; // invariants: // start < end // start < q <=end // array[start] < val && array[end] <=val // at every iteration d = end-start is decreasing // when d==0, then array[end] will be the first element >= val. if (array[q] >= val) { end = q; } else { start = q; } } return end; } /** * @brief Find the k-th occurrence of value in a sorted array. * * Assume that array is [0, 1, 1, 1, 2, 2, 4, 4, 4, 4, 6, 7] * then find_nth_occurrence(cset, 12, 4, 2) == 7, because cset_array[7] stores * the second element with value = 4. * If there are less than k values in the array, then return -1 * * @param [in] array sorted array of numbers, size [n] * @param [in] n number of elements in the array * @param [in] val the value we are searching for * @param [in] k * @return the idx of the k-th occurrence of val in array, or -1 if * the value is not found. */ int DI find_nth_occurrence(const int* array, int n, int val, int k) { int q = arg_first_ge(array, n, val); if (q + k < n && array[q + k] == val) { q += k; } else { q = -1; } return q; } /** * @brief Rank the entries in a cache set according to the time stamp, return * the indices that would sort the time stamp in ascending order. * * Assume we have a single cache set with time stamps as: * key (threadIdx.x): 0 1 2 3 * val (time stamp): 8 6 7 5 * * The corresponding sorted key-value pairs: * key: 3 1 2 0 * val: 5 6 7 8 * rank: 0th 1st 2nd 3rd * * On return, the rank is assigned for each thread: * threadIdx.x: 0 1 2 3 * rank: 3 1 2 0 * * For multiple cache sets, launch one block per cache set. * * @tparam nthreads number of threads per block (nthreads <= associativity) * @tparam associativity number of items in a cache set * * @param [in] cache_time time stamp of caching the data, size [associativity * n_cache_sets] * @param [in] n_cache_sets number of cache sets * @param [out] rank within the cache set size [nthreads * items_per_thread] * Each block should give a different pointer for rank. */ template <int nthreads, int associativity> DI void rank_set_entries(const int* cache_time, int n_cache_sets, int* rank) { const int items_per_thread = raft::ceildiv(associativity, nthreads); typedef cub::BlockRadixSort<int, nthreads, items_per_thread, int> BlockRadixSort; __shared__ typename BlockRadixSort::TempStorage temp_storage; int key[items_per_thread]; int val[items_per_thread]; int block_offset = blockIdx.x * associativity; for (int j = 0; j < items_per_thread; j++) { int k = threadIdx.x + j * nthreads; int t = (k < associativity) ? cache_time[block_offset + k] : 32768; key[j] = t; val[j] = k; } BlockRadixSort(temp_storage).Sort(key, val); for (int j = 0; j < items_per_thread; j++) { if (val[j] < associativity) { rank[val[j]] = threadIdx.x * items_per_thread + j; } } __syncthreads(); } /** * @brief Assign cache location to a set of keys using LRU replacement policy. * * The keys and the corresponding cache_set arrays shall be sorted according * to cache_set in ascending order. One block should be launched for every cache * set. * * Each cache set is sorted according to time_stamp, and values from keys * are filled in starting at the oldest time stamp. Entries that were accessed * at the current time are not reassigned. * * @tparam nthreads number of threads per block * @tparam associativity number of keys in a cache set * * @param [in] keys that we want to cache size [n] * @param [in] n number of keys * @param [in] cache_set assigned to keys, size [n] * @param [inout] cached_keys keys of already cached vectors, * size [n_cache_sets*associativity], on exit it will be updated with the * cached elements from keys. * @param [in] n_cache_sets number of cache sets * @param [inout] cache_time will be updated to "time" for those elements that * could be assigned to a cache location, size [n_cache_sets*associativity] * @param [in] time time stamp * @param [out] cache_idx the cache idx assigned to the input, or -1 if it could * not be cached, size [n] */ template <int nthreads, int associativity> RAFT_KERNEL assign_cache_idx(const int* keys, int n, const int* cache_set, int* cached_keys, int n_cache_sets, int* cache_time, int time, int* cache_idx) { int block_offset = blockIdx.x * associativity; const int items_per_thread = raft::ceildiv(associativity, nthreads); // the size of rank limits how large associativity can be used in practice __shared__ int rank[items_per_thread * nthreads]; rank_set_entries<nthreads, associativity>(cache_time, n_cache_sets, rank); // Each thread will fill items_per_thread items in the cache. // It uses a place, only if it was not updated at the current time step // (cache_time != time). // We rank the places according to the time stamp, least recently used // elements come to the front. // We fill the least recently used elements with the working set. // there might be elements which cannot be assigned to cache loc. // these elements are assigned -1. for (int j = 0; j < items_per_thread; j++) { int i = threadIdx.x + j * nthreads; int t_idx = block_offset + i; bool mask = (i < associativity); // whether this slot is available for writing mask = mask && (cache_time[t_idx] != time); // rank[i] tells which element to store by this thread // we look up where is the corresponding key stored in the input array if (mask) { int k = find_nth_occurrence(cache_set, n, blockIdx.x, rank[i]); if (k > -1) { int key_val = keys[k]; cached_keys[t_idx] = key_val; cache_idx[k] = t_idx; cache_time[t_idx] = time; } } } } /** * @brief Get the cache indices for keys stored in the cache. * * For every key, we look up the corresponding cache position. * If keys[k] is stored in the cache, then is_cached[k] is set to true, and * cache_idx[k] stores the corresponding cache idx. * * If keys[k] is not stored in the cache, then we assign a cache set to it. * This cache set is stored in cache_idx[k], and is_cached[k] is set to false. * In this case AssignCacheIdx should be called, to get an assigned position * within the cache set. * * Cache_time is assigned to the time input argument for all elements in idx. * * @param [in] keys array of keys that we want to look up in the cache, size [n] * @param [in] n number of keys to look up * @param [inout] cached_keys keys stored in the cache, size [n_cache_sets * associativity] * @param [in] n_cache_sets number of cache sets * @param [in] associativity number of keys in cache set * @param [inout] cache_time time stamp when the indices were cached, size [n_cache_sets * * associativity] * @param [out] cache_idx cache indices of the working set elements, size [n] * @param [out] is_cached whether the element is cached size[n] * @param [in] time iteration counter (used for time stamping) */ template <typename = void> RAFT_KERNEL get_cache_idx(int* keys, int n, int* cached_keys, int n_cache_sets, int associativity, int* cache_time, int* cache_idx, bool* is_cached, int time) { int tid = threadIdx.x + blockIdx.x * blockDim.x; if (tid < n) { int widx = keys[tid]; int sidx = hash(widx, n_cache_sets); int cidx = sidx * associativity; int i = 0; bool found = false; // search for empty spot and the least recently used spot while (i < associativity && !found) { found = (cache_time[cidx + i] > 0 && cached_keys[cidx + i] == widx); i++; } is_cached[tid] = found; if (found) { cidx = cidx + i - 1; cache_time[cidx] = time; // update time stamp cache_idx[tid] = cidx; // exact cache idx } else { cache_idx[tid] = sidx; // assign cache set } } } }; // namespace cache }; // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/memory_pool.hpp
/* * Copyright (c) 2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include "memory_pool-ext.hpp" #if !defined(RAFT_COMPILED) #include "memory_pool-inl.hpp" #endif // RAFT_COMPILED
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/itertools.hpp
/* * Copyright (c) 2022, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/util/detail/itertools.hpp> /** * Helpers inspired by the Python itertools library * */ namespace raft::util::itertools { /** * @brief Cartesian product of the given initializer lists. * * This helper can be used to easily define input parameters in tests/benchmarks. * Note that it's not optimized for use with large lists / many lists in performance-critical code! * * @tparam S Type of the output structures. * @tparam Args Types of the elements of the initilizer lists, matching the types of the first * fields of the structure (if the structure has more fields, some might be initialized * with their default value). * @param lists One or more initializer lists. * @return std::vector<S> A vector of structures containing the cartesian product. */ template <typename S, typename... Args> std::vector<S> product(std::initializer_list<Args>... lists) { return detail::product<S>(std::index_sequence_for<Args...>(), (std::vector<Args>(lists))...); } } // namespace raft::util::itertools
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/cache.cuh
/* * Copyright (c) 2019-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <cub/cub.cuh> #include <raft/core/interruptible.hpp> #include <raft/core/logger.hpp> #include <raft/util/cache_util.cuh> #include <raft/util/cuda_utils.cuh> #include <raft/util/cudart_utils.hpp> #include <rmm/device_scalar.hpp> #include <rmm/device_uvector.hpp> #include <cstddef> namespace raft::cache { /** * @brief Associative cache with least recently used replacement policy. * * SW managed cache in device memory, for ML algos where we can trade memory * access for computation. The two main functions of this class are the * management of cache indices, and methods to retrieve/store data using the * cache indices. * * The index management can be considered as a hash map<int, int>, where the int * keys are the original vector indices that we want to store, and the values are * the cache location of these vectors. The keys are hashed into a bucket * whose size equals the associativity. These are the cache sets. If a cache * set is full, then new indices are stored by replacing the oldest entries. * * Using this index mapping we implement methods to store and retrieve data from * the cache buffer, where a unit of data that we are storing is math_t[n_vec]. * For example in SVM we store full columns of the kernel matrix at each cache * entry. * * Note: we should have a look if the index management could be simplified using * concurrent_unordered_map.cuh from cudf. See Issue #914. * * Example usage: * @code{.cpp} * * // An expensive calculation that we want to accelerate with caching: * // we have n keys, and for each key we generate a vector with m elements. * // The keys and the output values are stored in GPU memory. * void calc(int *key, int n, int m, float *out, cudaStream_t stream) { * for (k=0; k<n; k++) { * // use key[k] to generate out[i + m*k], where i=0..m-1 * } * } * * // We assume that our ML algo repeatedly calls calc, and the set of keys have * // an overlap. We will use the cache to avoid repeated calculations. * * // Assume we have raft::resources& h, and cudaStream_t stream * Cache<float> cache(h.get_device_allocator(), stream, m); * * // A buffer that we will reuse to store the cache indices. * rmm::device_uvector<int> cache_idx(h.get_device_allocator(), stream, n); * * void cached_calc(int *key, int n, int m, float *out, stream) { * int n_cached = 0; * * cache.GetCacheIdxPartitioned(key, n, cache_idx.data(), &n_cached, * cudaStream_t stream); * * // Note: GetCacheIdxPartitioned has reordered the keys so that * // key[0..n_cached-1] are the keys already in the cache. * // We collect the corresponding values * cache.GetVecs(cache_idx.data(), n_cached, out, stream); * * // Calculate the elements not in the cache * int non_cached = n - n_cached; * if (non_cached > 0) { * int *key_new = key + n_cached; * int *cache_idx_new = cache_idx.data() + n_cached; * float *out_new = out + n_cached * m; * // AssignCacheIdx can permute the keys, therefore it has to come before * // we call calc. * // Note: a call to AssignCacheIdx should always be preceded with * // GetCacheIdxPartitioned, because that initializes the cache_idx_new array * // with the cache set (hash bucket) that correspond to the keys. * // The cache idx will be assigned from that cache set. * cache.AssignCacheIdx(key_new, non_cached, cache_idx_new, stream); * * calc(key_new, non_cached, m, out_new, stream); * * // Store the calculated vectors into the cache. * cache.StoreVecs(out_new, non_cached, non_cached, cache_idx_new, stream); * } * } * @endcode */ template <typename math_t, int associativity = 32> class Cache { public: /** * @brief Construct a Cache object * * @tparam math_t type of elements to be cached * @tparam associativity number of vectors in a cache set * * @param stream cuda stream * @param n_vec number of elements in a single vector that is stored in a * cache entry * @param cache_size in MiB */ Cache(cudaStream_t stream, int n_vec, float cache_size = 200) : n_vec(n_vec), cache_size(cache_size), cache(0, stream), cached_keys(0, stream), cache_time(0, stream), is_cached(0, stream), ws_tmp(0, stream), idx_tmp(0, stream), d_num_selected_out(stream), d_temp_storage(0, stream) { ASSERT(n_vec > 0, "Parameter n_vec: shall be larger than zero"); ASSERT(associativity > 0, "Associativity shall be larger than zero"); ASSERT(cache_size >= 0, "Cache size should not be negative"); // Calculate how many vectors would fit the cache int n_cache_vecs = (cache_size * 1024 * 1024) / (sizeof(math_t) * n_vec); // The available memory shall be enough for at least one cache set if (n_cache_vecs >= associativity) { n_cache_sets = n_cache_vecs / associativity; n_cache_vecs = n_cache_sets * associativity; cache.resize(n_cache_vecs * n_vec, stream); cached_keys.resize(n_cache_vecs, stream); cache_time.resize(n_cache_vecs, stream); RAFT_CUDA_TRY( cudaMemsetAsync(cached_keys.data(), 0, cached_keys.size() * sizeof(int), stream)); RAFT_CUDA_TRY(cudaMemsetAsync(cache_time.data(), 0, cache_time.size() * sizeof(int), stream)); } else { if (cache_size > 0) { RAFT_LOG_WARN( "Warning: not enough memory to cache a single set of " "rows, not using cache"); } n_cache_sets = 0; cache_size = 0; } RAFT_LOG_DEBUG( "Creating cache with size=%f MiB, to store %d vectors, in " "%d sets with associativity=%d", cache_size, n_cache_vecs, n_cache_sets, associativity); } Cache(const Cache& other) = delete; Cache& operator=(const Cache& other) = delete; /** @brief Collect cached data into contiguous memory space. * * On exit, the tile array is filled the following way: * out[i + n_vec*k] = cache[i + n_vec * idx[k]]), where i=0..n_vec-1, * k = 0..n-1 * * Idx values less than 0 are ignored. * * @param [in] idx cache indices, size [n] * @param [in] n the number of vectors that need to be collected * @param [out] out vectors collected from cache, size [n_vec*n] * @param [in] stream cuda stream */ void GetVecs(const int* idx, int n, math_t* out, cudaStream_t stream) { if (n > 0) { get_vecs<<<raft::ceildiv(n * n_vec, TPB), TPB, 0, stream>>>(cache.data(), n_vec, idx, n, out); RAFT_CUDA_TRY(cudaPeekAtLastError()); } } /** @brief Store vectors of data into the cache. * * Roughly the opposite of GetVecs, but the input vectors can be scattered * in memory. The cache is updated using the following formula: * * cache[i + cache_idx[k]*n_vec] = tile[i + tile_idx[k]*n_vec], * for i=0..n_vec-1, k=0..n-1 * * If tile_idx==nullptr, then we assume tile_idx[k] = k. * * Elements within a vector should be contiguous in memory (i.e. column vectors * for column major data storage, or row vectors of row major data). * * @param [in] tile stores the data to be cashed cached, size [n_vec x n_tile] * @param [in] n_tile number of vectors in tile (at least n) * @param [in] n number of vectors that need to be stored in the cache (a subset * of all the vectors in the tile) * @param [in] cache_idx cache indices for storing the vectors (negative values * are ignored), size [n] * @param [in] stream cuda stream * @param [in] tile_idx indices of vectors that need to be stored */ void StoreVecs(const math_t* tile, int n_tile, int n, int* cache_idx, cudaStream_t stream, const int* tile_idx = nullptr) { if (n > 0) { store_vecs<<<raft::ceildiv(n * n_vec, TPB), TPB, 0, stream>>>( tile, n_tile, n_vec, tile_idx, n, cache_idx, cache.data(), cache.size() / n_vec); RAFT_CUDA_TRY(cudaPeekAtLastError()); } } /** @brief Map a set of keys to cache indices. * * For each k in 0..n-1, if keys[k] is found in the cache, then cache_idx[k] * will tell the corresponding cache idx, and is_cached[k] is set to true. * * If keys[k] is not found in the cache, then is_cached[k] is set to false. * In this case we assign the cache set for keys[k], and cache_idx[k] will * store the cache set. * * @note in order to retrieve the cached vector j=cache_idx[k] from the cache, * we have to access cache[i + j*n_vec], where i=0..n_vec-1. * * @note: do not use simultaneous GetCacheIdx and AssignCacheIdx * * @param [in] keys device array of keys, size [n] * @param [in] n number of keys * @param [out] cache_idx device array of cache indices corresponding to the * input keys, size [n] * @param [out] is_cached whether the element is already available in the * cache, size [n] * @param [in] stream */ void GetCacheIdx(int* keys, int n, int* cache_idx, bool* is_cached, cudaStream_t stream) { n_iter++; // we increase the iteration counter, that is used to time stamp // accessing entries from the cache get_cache_idx<<<raft::ceildiv(n, TPB), TPB, 0, stream>>>(keys, n, cached_keys.data(), n_cache_sets, associativity, cache_time.data(), cache_idx, is_cached, n_iter); RAFT_CUDA_TRY(cudaPeekAtLastError()); } /** @brief Map a set of keys to cache indices. * * Same as GetCacheIdx, but partitions the keys, and cache_idx arrays in a way * that keys[0..n_cached-1] and cache_idx[0..n_cached-1] store the indices of * vectors that are found in the cache, while keys[n_cached..n-1] are the * indices of vectors that are not found in the cache. For the vectors not * found in the cache, cache_idx[n_cached..n-1] stores the cache set, and this * can be used to call AssignCacheIdx. * * @param [inout] keys device array of keys, size [n] * @param [in] n number of indices * @param [out] cache_idx device array of cache indices corresponding to * the input keys, size [n] * @param [out] n_cached number of elements that are cached * @param [in] stream cuda stream */ void GetCacheIdxPartitioned(int* keys, int n, int* cache_idx, int* n_cached, cudaStream_t stream) { ResizeTmpBuffers(n, stream); GetCacheIdx(keys, n, ws_tmp.data(), is_cached.data(), stream); // Group cache indices as [already cached, non_cached] cub::DevicePartition::Flagged(d_temp_storage.data(), d_temp_storage_size, ws_tmp.data(), is_cached.data(), cache_idx, d_num_selected_out.data(), n, stream); raft::update_host(n_cached, d_num_selected_out.data(), 1, stream); // Similarly re-group the input indices raft::copy(ws_tmp.data(), keys, n, stream); cub::DevicePartition::Flagged(d_temp_storage.data(), d_temp_storage_size, ws_tmp.data(), is_cached.data(), keys, d_num_selected_out.data(), n, stream); raft::interruptible::synchronize(stream); } /** * @brief Assign cache location to a set of keys. * * Note: call GetCacheIdx first, to get the cache_set assigned to the keys. * Keys that cannot be cached are assigned to -1. * * @param [inout] keys device array of keys, size [n] * @param [in] n number of elements that we want to cache * @param [inout] cidx on entry: cache_set, on exit: assigned cache_idx or -1, * size[n] * @param [in] stream cuda stream */ void AssignCacheIdx(int* keys, int n, int* cidx, cudaStream_t stream) { if (n <= 0) return; cub::DeviceRadixSort::SortPairs(d_temp_storage.data(), d_temp_storage_size, cidx, ws_tmp.data(), keys, idx_tmp.data(), n, 0, sizeof(int) * 8, stream); raft::copy(keys, idx_tmp.data(), n, stream); // set it to -1 RAFT_CUDA_TRY(cudaMemsetAsync(cidx, 255, n * sizeof(int), stream)); const int nthreads = associativity <= 32 ? associativity : 32; assign_cache_idx<nthreads, associativity><<<n_cache_sets, nthreads, 0, stream>>>( keys, n, ws_tmp.data(), cached_keys.data(), n_cache_sets, cache_time.data(), n_iter, cidx); RAFT_CUDA_TRY(cudaPeekAtLastError()); if (debug_mode) RAFT_CUDA_TRY(cudaDeviceSynchronize()); } /** Return approximate cache size in MiB. */ float GetSizeInMiB() const { return cache_size; } /** * Returns the number of vectors that can be cached. */ int GetSize() const { return cached_keys.size(); } protected: int n_vec; //!< Number of elements in a cached vector float cache_size; //!< in MiB int n_cache_sets; //!< number of cache sets const int TPB = 256; //!< threads per block for kernel launch int n_iter = 0; //!< Counter for time stamping cache operation bool debug_mode = false; rmm::device_uvector<math_t> cache; //!< The value of cached vectors rmm::device_uvector<int> cached_keys; //!< Keys stored at each cache loc rmm::device_uvector<int> cache_time; //!< Time stamp for LRU cache // Helper arrays for GetCacheIdx rmm::device_uvector<bool> is_cached; rmm::device_uvector<int> ws_tmp; rmm::device_uvector<int> idx_tmp; // Helper arrays for cub rmm::device_scalar<int> d_num_selected_out; rmm::device_uvector<char> d_temp_storage; size_t d_temp_storage_size = 0; void ResizeTmpBuffers(int n, cudaStream_t stream) { if (ws_tmp.size() < static_cast<std::size_t>(n)) { ws_tmp.resize(n, stream); is_cached.resize(n, stream); idx_tmp.resize(n, stream); cub::DevicePartition::Flagged(NULL, d_temp_storage_size, cached_keys.data(), is_cached.data(), cached_keys.data(), d_num_selected_out.data(), n, stream); d_temp_storage.resize(d_temp_storage_size, stream); } } }; }; // namespace raft::cache
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/cuda_utils.cuh
/* * Copyright (c) 2018-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <math_constants.h> #include <stdint.h> #include <type_traits> #if defined(_RAFT_HAS_CUDA) #include <cuda_bf16.h> #include <cuda_fp16.h> #endif #include <raft/core/cudart_utils.hpp> #include <raft/core/math.hpp> #include <raft/core/operators.hpp> // For backward compatibility, we include the follow headers. They contain // functionality that were previously contained in cuda_utils.cuh #include <raft/util/cuda_dev_essentials.cuh> #include <raft/util/reduction.cuh> namespace raft { /** Device function to have atomic add support for older archs */ template <typename Type> DI void myAtomicAdd(Type* address, Type val) { atomicAdd(address, val); } #if defined(__CUDA_ARCH__) && (__CUDA_ARCH__ < 600) // Ref: // http://on-demand.gputechconf.com/gtc/2013/presentations/S3101-Atomic-Memory-Operations.pdf template <> DI void myAtomicAdd(double* address, double val) { unsigned long long int* address_as_ull = (unsigned long long int*)address; unsigned long long int old = *address_as_ull, assumed; do { assumed = old; old = atomicCAS(address_as_ull, assumed, __double_as_longlong(val + __longlong_as_double(assumed))); } while (assumed != old); } #endif template <typename T, typename ReduceLambda> DI void myAtomicReduce(T* address, T val, ReduceLambda op); template <typename ReduceLambda> DI void myAtomicReduce(double* address, double val, ReduceLambda op) { unsigned long long int* address_as_ull = (unsigned long long int*)address; unsigned long long int old = *address_as_ull, assumed; do { assumed = old; old = atomicCAS( address_as_ull, assumed, __double_as_longlong(op(val, __longlong_as_double(assumed)))); } while (assumed != old); } template <typename ReduceLambda> DI void myAtomicReduce(float* address, float val, ReduceLambda op) { unsigned int* address_as_uint = (unsigned int*)address; unsigned int old = *address_as_uint, assumed; do { assumed = old; old = atomicCAS(address_as_uint, assumed, __float_as_uint(op(val, __uint_as_float(assumed)))); } while (assumed != old); } // Needed for atomicCas on ushort #if defined(__CUDA_ARCH__) && (__CUDA_ARCH__ >= 700) template <typename ReduceLambda> DI void myAtomicReduce(__half* address, __half val, ReduceLambda op) { unsigned short int* address_as_uint = (unsigned short int*)address; unsigned short int old = *address_as_uint, assumed; do { assumed = old; old = atomicCAS(address_as_uint, assumed, __half_as_ushort(op(val, __ushort_as_half(assumed)))); } while (assumed != old); } #endif // Needed for nv_bfloat16 support #if defined(__CUDA_ARCH__) && (__CUDA_ARCH__ >= 800) template <typename ReduceLambda> DI void myAtomicReduce(nv_bfloat16* address, nv_bfloat16 val, ReduceLambda op) { unsigned short int* address_as_uint = (unsigned short int*)address; unsigned short int old = *address_as_uint, assumed; do { assumed = old; old = atomicCAS( address_as_uint, assumed, __bfloat16_as_ushort(op(val, __ushort_as_bfloat16(assumed)))); } while (assumed != old); } #endif template <typename ReduceLambda> DI void myAtomicReduce(int* address, int val, ReduceLambda op) { int old = *address, assumed; do { assumed = old; old = atomicCAS(address, assumed, op(val, assumed)); } while (assumed != old); } template <typename ReduceLambda> DI void myAtomicReduce(long long* address, long long val, ReduceLambda op) { long long old = *address, assumed; do { assumed = old; old = atomicCAS(address, assumed, op(val, assumed)); } while (assumed != old); } template <typename ReduceLambda> DI void myAtomicReduce(unsigned long long* address, unsigned long long val, ReduceLambda op) { unsigned long long old = *address, assumed; do { assumed = old; old = atomicCAS(address, assumed, op(val, assumed)); } while (assumed != old); } /** * @brief Provide atomic min operation. * @tparam T: data type for input data (float or double). * @param[in] address: address to read old value from, and to atomically update w/ min(old value, * val) * @param[in] val: new value to compare with old */ template <typename T> DI T myAtomicMin(T* address, T val); /** * @brief Provide atomic max operation. * @tparam T: data type for input data (float or double). * @param[in] address: address to read old value from, and to atomically update w/ max(old value, * val) * @param[in] val: new value to compare with old */ template <typename T> DI T myAtomicMax(T* address, T val); DI float myAtomicMin(float* address, float val) { myAtomicReduce<float(float, float)>(address, val, fminf); return *address; } DI float myAtomicMax(float* address, float val) { myAtomicReduce<float(float, float)>(address, val, fmaxf); return *address; } DI double myAtomicMin(double* address, double val) { myAtomicReduce<double(double, double)>(address, val, fmin); return *address; } DI double myAtomicMax(double* address, double val) { myAtomicReduce<double(double, double)>(address, val, fmax); return *address; } /** * @defgroup Max maximum of two numbers * @{ */ template <typename T> HDI T myMax(T x, T y); template <> [[deprecated("use raft::max from raft/core/math.hpp instead")]] HDI float myMax<float>(float x, float y) { return fmaxf(x, y); } template <> [[deprecated("use raft::max from raft/core/math.hpp instead")]] HDI double myMax<double>(double x, double y) { return fmax(x, y); } /** @} */ /** * @defgroup Min minimum of two numbers * @{ */ template <typename T> HDI T myMin(T x, T y); template <> [[deprecated("use raft::min from raft/core/math.hpp instead")]] HDI float myMin<float>(float x, float y) { return fminf(x, y); } template <> [[deprecated("use raft::min from raft/core/math.hpp instead")]] HDI double myMin<double>(double x, double y) { return fmin(x, y); } /** @} */ /** * @brief Provide atomic min operation. * @tparam T: data type for input data (float or double). * @param[in] address: address to read old value from, and to atomically update w/ min(old value, * val) * @param[in] val: new value to compare with old */ template <typename T> DI T myAtomicMin(T* address, T val) { myAtomicReduce(address, val, raft::min_op{}); return *address; } /** * @brief Provide atomic max operation. * @tparam T: data type for input data (float or double). * @param[in] address: address to read old value from, and to atomically update w/ max(old value, * val) * @param[in] val: new value to compare with old */ template <typename T> DI T myAtomicMax(T* address, T val) { myAtomicReduce(address, val, raft::max_op{}); return *address; } /** * @defgroup Exp Exponential function * @{ */ template <typename T> HDI T myExp(T x); template <> [[deprecated("use raft::exp from raft/core/math.hpp instead")]] HDI float myExp(float x) { return expf(x); } template <> [[deprecated("use raft::exp from raft/core/math.hpp instead")]] HDI double myExp(double x) { return ::exp(x); } /** @} */ /** * @defgroup Cuda infinity values * @{ */ template <typename T> inline __device__ T myInf(); template <> inline __device__ float myInf<float>() { return CUDART_INF_F; } template <> inline __device__ double myInf<double>() { return CUDART_INF; } /** @} */ /** * @defgroup Log Natural logarithm * @{ */ template <typename T> HDI T myLog(T x); template <> [[deprecated("use raft::log from raft/core/math.hpp instead")]] HDI float myLog(float x) { return logf(x); } template <> [[deprecated("use raft::log from raft/core/math.hpp instead")]] HDI double myLog(double x) { return ::log(x); } /** @} */ /** * @defgroup Sqrt Square root * @{ */ template <typename T> HDI T mySqrt(T x); template <> [[deprecated("use raft::sqrt from raft/core/math.hpp instead")]] HDI float mySqrt(float x) { return sqrtf(x); } template <> [[deprecated("use raft::sqrt from raft/core/math.hpp instead")]] HDI double mySqrt(double x) { return ::sqrt(x); } /** @} */ /** * @defgroup SineCosine Sine and cosine calculation * @{ */ template <typename T> DI void mySinCos(T x, T& s, T& c); template <> [[deprecated("use raft::sincos from raft/core/math.hpp instead")]] DI void mySinCos(float x, float& s, float& c) { sincosf(x, &s, &c); } template <> [[deprecated("use raft::sincos from raft/core/math.hpp instead")]] DI void mySinCos(double x, double& s, double& c) { ::sincos(x, &s, &c); } /** @} */ /** * @defgroup Sine Sine calculation * @{ */ template <typename T> DI T mySin(T x); template <> [[deprecated("use raft::sin from raft/core/math.hpp instead")]] DI float mySin(float x) { return sinf(x); } template <> [[deprecated("use raft::sin from raft/core/math.hpp instead")]] DI double mySin(double x) { return ::sin(x); } /** @} */ /** * @defgroup Abs Absolute value * @{ */ template <typename T> DI T myAbs(T x) { return x < 0 ? -x : x; } template <> [[deprecated("use raft::abs from raft/core/math.hpp instead")]] DI float myAbs(float x) { return fabsf(x); } template <> [[deprecated("use raft::abs from raft/core/math.hpp instead")]] DI double myAbs(double x) { return fabs(x); } /** @} */ /** * @defgroup Pow Power function * @{ */ template <typename T> HDI T myPow(T x, T power); template <> [[deprecated("use raft::pow from raft/core/math.hpp instead")]] HDI float myPow(float x, float power) { return powf(x, power); } template <> [[deprecated("use raft::pow from raft/core/math.hpp instead")]] HDI double myPow(double x, double power) { return ::pow(x, power); } /** @} */ /** * @defgroup myTanh tanh function * @{ */ template <typename T> HDI T myTanh(T x); template <> [[deprecated("use raft::tanh from raft/core/math.hpp instead")]] HDI float myTanh(float x) { return tanhf(x); } template <> [[deprecated("use raft::tanh from raft/core/math.hpp instead")]] HDI double myTanh(double x) { return ::tanh(x); } /** @} */ /** * @defgroup myATanh arctanh function * @{ */ template <typename T> HDI T myATanh(T x); template <> [[deprecated("use raft::atanh from raft/core/math.hpp instead")]] HDI float myATanh(float x) { return atanhf(x); } template <> [[deprecated("use raft::atanh from raft/core/math.hpp instead")]] HDI double myATanh(double x) { return ::atanh(x); } /** @} */ /** * @defgroup LambdaOps Legacy lambda operations, to be deprecated * @{ */ template <typename Type, typename IdxType = int> struct Nop { [[deprecated("Nop is deprecated. Use identity_op instead.")]] HDI Type operator()(Type in, IdxType i = 0) const { return in; } }; template <typename Type, typename IdxType = int> struct SqrtOp { [[deprecated("SqrtOp is deprecated. Use sqrt_op instead.")]] HDI Type operator()(Type in, IdxType i = 0) const { return raft::sqrt(in); } }; template <typename Type, typename IdxType = int> struct L0Op { [[deprecated("L0Op is deprecated. Use nz_op instead.")]] HDI Type operator()(Type in, IdxType i = 0) const { return in != Type(0) ? Type(1) : Type(0); } }; template <typename Type, typename IdxType = int> struct L1Op { [[deprecated("L1Op is deprecated. Use abs_op instead.")]] HDI Type operator()(Type in, IdxType i = 0) const { return raft::abs(in); } }; template <typename Type, typename IdxType = int> struct L2Op { [[deprecated("L2Op is deprecated. Use sq_op instead.")]] HDI Type operator()(Type in, IdxType i = 0) const { return in * in; } }; template <typename InT, typename OutT = InT> struct Sum { [[deprecated("Sum is deprecated. Use add_op instead.")]] HDI OutT operator()(InT a, InT b) const { return a + b; } }; template <typename Type> struct Max { [[deprecated("Max is deprecated. Use max_op instead.")]] HDI Type operator()(Type a, Type b) const { if (b > a) { return b; } return a; } }; /** @} */ /** * @defgroup Sign Obtain sign value * @brief Obtain sign of x * @param x input * @return +1 if x >= 0 and -1 otherwise * @{ */ template <typename T> DI T signPrim(T x) { return x < 0 ? -1 : +1; } template <> DI float signPrim(float x) { return signbit(x) == true ? -1.0f : +1.0f; } template <> DI double signPrim(double x) { return signbit(x) == true ? -1.0 : +1.0; } /** @} */ /** * @defgroup Max maximum of two numbers * @brief Obtain maximum of two values * @param x one item * @param y second item * @return maximum of two items * @{ */ template <typename T> DI T maxPrim(T x, T y) { return x > y ? x : y; } template <> DI float maxPrim(float x, float y) { return fmaxf(x, y); } template <> DI double maxPrim(double x, double y) { return fmax(x, y); } /** @} */ /** * @brief Four-way byte dot product-accumulate. * @tparam T Four-byte integer: int or unsigned int * @tparam S Either same as T or a 4-byte vector of the same signedness. * * @param a * @param b * @param c * @return dot(a, b) + c */ template <typename T, typename S = T> DI auto dp4a(S a, S b, T c) -> T; template <> DI auto dp4a(char4 a, char4 b, int c) -> int { #if __CUDA_ARCH__ >= 610 return __dp4a(a, b, c); #else c += static_cast<int>(a.x) * static_cast<int>(b.x); c += static_cast<int>(a.y) * static_cast<int>(b.y); c += static_cast<int>(a.z) * static_cast<int>(b.z); c += static_cast<int>(a.w) * static_cast<int>(b.w); return c; #endif } template <> DI auto dp4a(uchar4 a, uchar4 b, unsigned int c) -> unsigned int { #if __CUDA_ARCH__ >= 610 return __dp4a(a, b, c); #else c += static_cast<unsigned int>(a.x) * static_cast<unsigned int>(b.x); c += static_cast<unsigned int>(a.y) * static_cast<unsigned int>(b.y); c += static_cast<unsigned int>(a.z) * static_cast<unsigned int>(b.z); c += static_cast<unsigned int>(a.w) * static_cast<unsigned int>(b.w); return c; #endif } template <> DI auto dp4a(int a, int b, int c) -> int { #if __CUDA_ARCH__ >= 610 return __dp4a(a, b, c); #else return dp4a(*reinterpret_cast<char4*>(&a), *reinterpret_cast<char4*>(&b), c); #endif } template <> DI auto dp4a(unsigned int a, unsigned int b, unsigned int c) -> unsigned int { #if __CUDA_ARCH__ >= 610 return __dp4a(a, b, c); #else return dp4a(*reinterpret_cast<uchar4*>(&a), *reinterpret_cast<uchar4*>(&b), c); #endif } /** * @brief Simple utility function to determine whether user_stream or one of the * internal streams should be used. * @param user_stream main user stream * @param int_streams array of internal streams * @param n_int_streams number of internal streams * @param idx the index for which to query the stream */ inline cudaStream_t select_stream(cudaStream_t user_stream, cudaStream_t* int_streams, int n_int_streams, int idx) { return n_int_streams > 0 ? int_streams[idx % n_int_streams] : user_stream; } } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/pow2_utils.cuh
/* * Copyright (c) 2021-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/util/cuda_utils.cuh> namespace raft { /** * @brief Fast arithmetics and alignment checks for power-of-two values known at compile time. * * @tparam Value_ a compile-time value representable as a power-of-two. */ template <auto Value_> struct Pow2 { typedef decltype(Value_) Type; static constexpr Type Value = Value_; static constexpr Type Log2 = log2(Value); static constexpr Type Mask = Value - 1; static_assert(std::is_integral<Type>::value, "Value must be integral."); static_assert(Value && !(Value & Mask), "Value must be power of two."); #define Pow2_FUNC_QUALIFIER static constexpr __host__ __device__ __forceinline__ #define Pow2_WHEN_INTEGRAL(I) std::enable_if_t<Pow2_IS_REPRESENTABLE_AS(I), I> #define Pow2_IS_REPRESENTABLE_AS(I) (std::is_integral<I>::value && Type(I(Value)) == Value) /** * Integer division by Value truncated toward zero * (same as `x / Value` in C++). * * Invariant: `x = Value * quot(x) + rem(x)` */ template <typename I> Pow2_FUNC_QUALIFIER Pow2_WHEN_INTEGRAL(I) quot(I x) noexcept { if constexpr (std::is_signed<I>::value) return (x >> I(Log2)) + (x < 0 && (x & I(Mask))); if constexpr (std::is_unsigned<I>::value) return x >> I(Log2); } /** * Remainder of integer division by Value truncated toward zero * (same as `x % Value` in C++). * * Invariant: `x = Value * quot(x) + rem(x)`. */ template <typename I> Pow2_FUNC_QUALIFIER Pow2_WHEN_INTEGRAL(I) rem(I x) noexcept { if constexpr (std::is_signed<I>::value) return x < 0 ? -((-x) & I(Mask)) : (x & I(Mask)); if constexpr (std::is_unsigned<I>::value) return x & I(Mask); } /** * Integer division by Value truncated toward negative infinity * (same as `x // Value` in Python). * * Invariant: `x = Value * div(x) + mod(x)`. * * Note, `div` and `mod` for negative values are slightly faster * than `quot` and `rem`, but behave slightly different * compared to normal C++ operators `/` and `%`. */ template <typename I> Pow2_FUNC_QUALIFIER Pow2_WHEN_INTEGRAL(I) div(I x) noexcept { return x >> I(Log2); } /** * Rounds up the value to next power of two. */ template <typename I> Pow2_FUNC_QUALIFIER Pow2_WHEN_INTEGRAL(I) round_up_pow2(I val) noexcept { return 1 << (log2(val) + 1); } /** * x modulo Value operation (remainder of the `div(x)`) * (same as `x % Value` in Python). * * Invariant: `mod(x) >= 0` * Invariant: `x = Value * div(x) + mod(x)`. * * Note, `div` and `mod` for negative values are slightly faster * than `quot` and `rem`, but behave slightly different * compared to normal C++ operators `/` and `%`. */ template <typename I> Pow2_FUNC_QUALIFIER Pow2_WHEN_INTEGRAL(I) mod(I x) noexcept { return x & I(Mask); } #define Pow2_CHECK_TYPE(T) \ static_assert(std::is_pointer<T>::value || std::is_integral<T>::value, \ "Only pointer or integral types make sense here") /** * Tell whether the pointer or integral is Value-aligned. * NB: for pointers, the alignment is checked in bytes, not in elements. */ template <typename PtrT> Pow2_FUNC_QUALIFIER bool isAligned(PtrT p) noexcept { Pow2_CHECK_TYPE(PtrT); if constexpr (Pow2_IS_REPRESENTABLE_AS(PtrT)) return mod(p) == 0; if constexpr (!Pow2_IS_REPRESENTABLE_AS(PtrT)) return mod(reinterpret_cast<Type>(p)) == 0; } /** Tell whether two pointers have the same address modulo Value. */ template <typename PtrT, typename PtrS> Pow2_FUNC_QUALIFIER bool areSameAlignOffsets(PtrT a, PtrS b) noexcept { Pow2_CHECK_TYPE(PtrT); Pow2_CHECK_TYPE(PtrS); Type x, y; if constexpr (Pow2_IS_REPRESENTABLE_AS(PtrT)) x = Type(mod(a)); else x = mod(reinterpret_cast<Type>(a)); if constexpr (Pow2_IS_REPRESENTABLE_AS(PtrS)) y = Type(mod(b)); else y = mod(reinterpret_cast<Type>(b)); return x == y; } /** Get this or next Value-aligned address (in bytes) or integral. */ template <typename PtrT> Pow2_FUNC_QUALIFIER PtrT roundUp(PtrT p) noexcept { Pow2_CHECK_TYPE(PtrT); if constexpr (Pow2_IS_REPRESENTABLE_AS(PtrT)) return (p + PtrT(Mask)) & PtrT(~Mask); if constexpr (!Pow2_IS_REPRESENTABLE_AS(PtrT)) { auto x = reinterpret_cast<Type>(p); return reinterpret_cast<PtrT>((x + Mask) & (~Mask)); } } /** Get this or previous Value-aligned address (in bytes) or integral. */ template <typename PtrT> Pow2_FUNC_QUALIFIER PtrT roundDown(PtrT p) noexcept { Pow2_CHECK_TYPE(PtrT); if constexpr (Pow2_IS_REPRESENTABLE_AS(PtrT)) return p & PtrT(~Mask); if constexpr (!Pow2_IS_REPRESENTABLE_AS(PtrT)) { auto x = reinterpret_cast<Type>(p); return reinterpret_cast<PtrT>(x & (~Mask)); } } #undef Pow2_CHECK_TYPE #undef Pow2_IS_REPRESENTABLE_AS #undef Pow2_FUNC_QUALIFIER #undef Pow2_WHEN_INTEGRAL }; }; // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/input_validation.hpp
/* * Copyright (c) 2022, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once namespace raft { template <class ElementType, class Extents, class Layout, class Accessor> constexpr bool is_row_or_column_major(mdspan<ElementType, Extents, Layout, Accessor> m) { return false; } template <class ElementType, class Extents, class Accessor> constexpr bool is_row_or_column_major(mdspan<ElementType, Extents, layout_left, Accessor> m) { return true; } template <class ElementType, class Extents, class Accessor> constexpr bool is_row_or_column_major(mdspan<ElementType, Extents, layout_right, Accessor> m) { return true; } template <class ElementType, class Extents, class Accessor> constexpr bool is_row_or_column_major(mdspan<ElementType, Extents, layout_stride, Accessor> m) { return m.is_exhaustive(); } template <class ElementType, class Extents, class Layout, class Accessor> constexpr bool is_row_major(mdspan<ElementType, Extents, Layout, Accessor> /* m */) { return false; } template <class ElementType, class Extents, class Accessor> constexpr bool is_row_major(mdspan<ElementType, Extents, layout_left, Accessor> /* m */) { return false; } template <class ElementType, class Extents, class Accessor> constexpr bool is_row_major(mdspan<ElementType, Extents, layout_right, Accessor> /* m */) { return true; } template <class ElementType, class Extents, class Accessor> constexpr bool is_row_major(mdspan<ElementType, Extents, layout_stride, Accessor> m) { return m.is_exhaustive() && m.stride(1) == typename Extents::index_type(1); } template <class ElementType, class Extents, class Layout, class Accessor> constexpr bool is_col_major(mdspan<ElementType, Extents, Layout, Accessor> /* m */) { return false; } template <class ElementType, class Extents, class Accessor> constexpr bool is_col_major(mdspan<ElementType, Extents, layout_left, Accessor> /* m */) { return true; } template <class ElementType, class Extents, class Accessor> constexpr bool is_col_major(mdspan<ElementType, Extents, layout_right, Accessor> /* m */) { return false; } template <class ElementType, class Extents, class Accessor> constexpr bool is_col_major(mdspan<ElementType, Extents, layout_stride, Accessor> m) { return m.is_exhaustive() && m.stride(0) == typename Extents::index_type(1); } template <class ElementType, class IndexType, size_t... Exts, class Layout, class Accessor> constexpr bool is_matrix_view( mdspan<ElementType, extents<IndexType, Exts...>, Layout, Accessor> /* m */) { return sizeof...(Exts) == 2; } template <class ElementType, class Extents> constexpr bool is_matrix_view(mdspan<ElementType, Extents> m) { return false; } template <class ElementType, class IndexType, size_t... Exts, class Layout, class Accessor> constexpr bool is_vector_view( mdspan<ElementType, extents<IndexType, Exts...>, Layout, Accessor> /* m */) { return sizeof...(Exts) == 1; } template <class ElementType, class Extents> constexpr bool is_vector_view(mdspan<ElementType, Extents> m) { return false; } template <class ElementType, class IndexType, size_t... Exts, class Layout, class Accessor> constexpr bool is_scalar_view( mdspan<ElementType, extents<IndexType, Exts...>, Layout, Accessor> /* m */) { return sizeof...(Exts) == 0; } template <class ElementType, class Extents> constexpr bool is_scalar_view(mdspan<ElementType, Extents> m) { return false; } }; // end namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/device_atomics.cuh
/* * Copyright (c) 2019-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once /** * @brief overloads for CUDA atomic operations * @file device_atomics.cuh * * Provides the overloads for arithmetic data types, where CUDA atomic operations are, `atomicAdd`, * `atomicMin`, `atomicMax`, and `atomicCAS`. * `atomicAnd`, `atomicOr`, `atomicXor` are also supported for integer data types. * Also provides `raft::genericAtomicOperation` which performs atomic operation with the given * binary operator. */ #include <cooperative_groups.h> #include <type_traits> namespace raft { namespace device_atomics { namespace detail { // ------------------------------------------------------------------------------------------------- // Binary operators /* @brief binary `sum` operator */ struct DeviceSum { template <typename T, typename std::enable_if_t<std::is_arithmetic<T>::value>* = nullptr> __device__ T operator()(const T& lhs, const T& rhs) { return lhs + rhs; } }; /* @brief binary `min` operator */ struct DeviceMin { template <typename T> __device__ T operator()(const T& lhs, const T& rhs) { return lhs < rhs ? lhs : rhs; } }; /* @brief binary `max` operator */ struct DeviceMax { template <typename T> __device__ T operator()(const T& lhs, const T& rhs) { return lhs > rhs ? lhs : rhs; } }; /* @brief binary `product` operator */ struct DeviceProduct { template <typename T, typename std::enable_if_t<std::is_arithmetic<T>::value>* = nullptr> __device__ T operator()(const T& lhs, const T& rhs) { return lhs * rhs; } }; /* @brief binary `and` operator */ struct DeviceAnd { template <typename T, typename std::enable_if_t<std::is_integral<T>::value>* = nullptr> __device__ T operator()(const T& lhs, const T& rhs) { return (lhs & rhs); } }; /* @brief binary `or` operator */ struct DeviceOr { template <typename T, typename std::enable_if_t<std::is_integral<T>::value>* = nullptr> __device__ T operator()(const T& lhs, const T& rhs) { return (lhs | rhs); } }; /* @brief binary `xor` operator */ struct DeviceXor { template <typename T, typename std::enable_if_t<std::is_integral<T>::value>* = nullptr> __device__ T operator()(const T& lhs, const T& rhs) { return (lhs ^ rhs); } }; // FIXME: remove this if C++17 is supported. // `static_assert` requires a string literal at C++14. #define errmsg_cast "size mismatch." template <typename T_output, typename T_input> __forceinline__ __device__ T_output type_reinterpret(T_input value) { static_assert(sizeof(T_output) == sizeof(T_input), "type_reinterpret for different size"); return *(reinterpret_cast<T_output*>(&value)); } // ------------------------------------------------------------------------------------------------- // the implementation of `genericAtomicOperation` template <typename T, typename Op, size_t N = sizeof(T)> struct genericAtomicOperationImpl; // single byte atomic operation template <typename T, typename Op> struct genericAtomicOperationImpl<T, Op, 1> { __forceinline__ __device__ T operator()(T* addr, T const& update_value, Op op) { using T_int = unsigned int; T_int* address_uint32 = reinterpret_cast<T_int*>(addr - (reinterpret_cast<size_t>(addr) & 3)); T_int shift = ((reinterpret_cast<size_t>(addr) & 3) * 8); T_int old = *address_uint32; T_int assumed; do { assumed = old; T target_value = T((old >> shift) & 0xff); uint8_t updating_value = type_reinterpret<uint8_t, T>(op(target_value, update_value)); T_int new_value = (old & ~(0x000000ff << shift)) | (T_int(updating_value) << shift); old = atomicCAS(address_uint32, assumed, new_value); } while (assumed != old); return T((old >> shift) & 0xff); } }; // 2 bytes atomic operation template <typename T, typename Op> struct genericAtomicOperationImpl<T, Op, 2> { __forceinline__ __device__ T operator()(T* addr, T const& update_value, Op op) { using T_int = unsigned int; bool is_32_align = (reinterpret_cast<size_t>(addr) & 2) ? false : true; T_int* address_uint32 = reinterpret_cast<T_int*>(reinterpret_cast<size_t>(addr) - (is_32_align ? 0 : 2)); T_int old = *address_uint32; T_int assumed; do { assumed = old; T target_value = (is_32_align) ? T(old & 0xffff) : T(old >> 16); uint16_t updating_value = type_reinterpret<uint16_t, T>(op(target_value, update_value)); T_int new_value = (is_32_align) ? (old & 0xffff0000) | updating_value : (old & 0xffff) | (T_int(updating_value) << 16); old = atomicCAS(address_uint32, assumed, new_value); } while (assumed != old); return (is_32_align) ? T(old & 0xffff) : T(old >> 16); ; } }; // 4 bytes atomic operation template <typename T, typename Op> struct genericAtomicOperationImpl<T, Op, 4> { __forceinline__ __device__ T operator()(T* addr, T const& update_value, Op op) { using T_int = unsigned int; T old_value = *addr; T assumed{old_value}; if constexpr (std::is_same<T, float>{} && (std::is_same<Op, DeviceMin>{})) { if (isnan(update_value)) { return old_value; } } do { assumed = old_value; const T new_value = op(old_value, update_value); T_int ret = atomicCAS(reinterpret_cast<T_int*>(addr), type_reinterpret<T_int, T>(assumed), type_reinterpret<T_int, T>(new_value)); old_value = type_reinterpret<T, T_int>(ret); } while (assumed != old_value); return old_value; } }; // 4 bytes fp32 atomic Max operation template <> struct genericAtomicOperationImpl<float, DeviceMax, 4> { using T = float; __forceinline__ __device__ T operator()(T* addr, T const& update_value, DeviceMax op) { if (isnan(update_value)) { return *addr; } T old = (update_value >= 0) ? __int_as_float(atomicMax((int*)addr, __float_as_int(update_value))) : __uint_as_float(atomicMin((unsigned int*)addr, __float_as_uint(update_value))); return old; } }; // 8 bytes atomic operation template <typename T, typename Op> struct genericAtomicOperationImpl<T, Op, 8> { __forceinline__ __device__ T operator()(T* addr, T const& update_value, Op op) { using T_int = unsigned long long int; static_assert(sizeof(T) == sizeof(T_int), errmsg_cast); T old_value = *addr; T assumed{old_value}; do { assumed = old_value; const T new_value = op(old_value, update_value); T_int ret = atomicCAS(reinterpret_cast<T_int*>(addr), type_reinterpret<T_int, T>(assumed), type_reinterpret<T_int, T>(new_value)); old_value = type_reinterpret<T, T_int>(ret); } while (assumed != old_value); return old_value; } }; // ------------------------------------------------------------------------------------------------- // specialized functions for operators // `atomicAdd` supports int, unsigned int, unsigned long long int, float, double (long long int is // not supported.) `atomicMin`, `atomicMax` support int, unsigned int, unsigned long long int // `atomicAnd`, `atomicOr`, `atomicXor` support int, unsigned int, unsigned long long int // CUDA natively supports `unsigned long long int` for `atomicAdd`, // but doesn't supports `long int`. // However, since the signed integer is represented as Two's complement, // the fundamental arithmetic operations of addition are identical to // those for unsigned binary numbers. // Then, this computes as `unsigned long long int` with `atomicAdd` // @sa https://en.wikipedia.org/wiki/Two%27s_complement template <> struct genericAtomicOperationImpl<long int, DeviceSum, 8> { using T = long int; __forceinline__ __device__ T operator()(T* addr, T const& update_value, DeviceSum op) { using T_int = unsigned long long int; static_assert(sizeof(T) == sizeof(T_int), errmsg_cast); T_int ret = atomicAdd(reinterpret_cast<T_int*>(addr), type_reinterpret<T_int, T>(update_value)); return type_reinterpret<T, T_int>(ret); } }; template <> struct genericAtomicOperationImpl<unsigned long int, DeviceSum, 8> { using T = unsigned long int; __forceinline__ __device__ T operator()(T* addr, T const& update_value, DeviceSum op) { using T_int = unsigned long long int; static_assert(sizeof(T) == sizeof(T_int), errmsg_cast); T_int ret = atomicAdd(reinterpret_cast<T_int*>(addr), type_reinterpret<T_int, T>(update_value)); return type_reinterpret<T, T_int>(ret); } }; // CUDA natively supports `unsigned long long int` for `atomicAdd`, // but doesn't supports `long long int`. // However, since the signed integer is represented as Two's complement, // the fundamental arithmetic operations of addition are identical to // those for unsigned binary numbers. // Then, this computes as `unsigned long long int` with `atomicAdd` // @sa https://en.wikipedia.org/wiki/Two%27s_complement template <> struct genericAtomicOperationImpl<long long int, DeviceSum, 8> { using T = long long int; __forceinline__ __device__ T operator()(T* addr, T const& update_value, DeviceSum op) { using T_int = unsigned long long int; static_assert(sizeof(T) == sizeof(T_int), errmsg_cast); T_int ret = atomicAdd(reinterpret_cast<T_int*>(addr), type_reinterpret<T_int, T>(update_value)); return type_reinterpret<T, T_int>(ret); } }; template <> struct genericAtomicOperationImpl<unsigned long int, DeviceMin, 8> { using T = unsigned long int; __forceinline__ __device__ T operator()(T* addr, T const& update_value, DeviceMin op) { using T_int = unsigned long long int; static_assert(sizeof(T) == sizeof(T_int), errmsg_cast); T ret = atomicMin(reinterpret_cast<T_int*>(addr), type_reinterpret<T_int, T>(update_value)); return type_reinterpret<T, T_int>(ret); } }; template <> struct genericAtomicOperationImpl<unsigned long int, DeviceMax, 8> { using T = unsigned long int; __forceinline__ __device__ T operator()(T* addr, T const& update_value, DeviceMax op) { using T_int = unsigned long long int; static_assert(sizeof(T) == sizeof(T_int), errmsg_cast); T ret = atomicMax(reinterpret_cast<T_int*>(addr), type_reinterpret<T_int, T>(update_value)); return type_reinterpret<T, T_int>(ret); } }; template <typename T> struct genericAtomicOperationImpl<T, DeviceAnd, 8> { __forceinline__ __device__ T operator()(T* addr, T const& update_value, DeviceAnd op) { using T_int = unsigned long long int; static_assert(sizeof(T) == sizeof(T_int), errmsg_cast); T_int ret = atomicAnd(reinterpret_cast<T_int*>(addr), type_reinterpret<T_int, T>(update_value)); return type_reinterpret<T, T_int>(ret); } }; template <typename T> struct genericAtomicOperationImpl<T, DeviceOr, 8> { __forceinline__ __device__ T operator()(T* addr, T const& update_value, DeviceOr op) { using T_int = unsigned long long int; static_assert(sizeof(T) == sizeof(T_int), errmsg_cast); T_int ret = atomicOr(reinterpret_cast<T_int*>(addr), type_reinterpret<T_int, T>(update_value)); return type_reinterpret<T, T_int>(ret); } }; template <typename T> struct genericAtomicOperationImpl<T, DeviceXor, 8> { __forceinline__ __device__ T operator()(T* addr, T const& update_value, DeviceXor op) { using T_int = unsigned long long int; static_assert(sizeof(T) == sizeof(T_int), errmsg_cast); T_int ret = atomicXor(reinterpret_cast<T_int*>(addr), type_reinterpret<T_int, T>(update_value)); return type_reinterpret<T, T_int>(ret); } }; // ------------------------------------------------------------------------------------------------- // the implementation of `typesAtomicCASImpl` template <typename T, size_t N = sizeof(T)> struct typesAtomicCASImpl; template <typename T> struct typesAtomicCASImpl<T, 1> { __forceinline__ __device__ T operator()(T* addr, T const& compare, T const& update_value) { using T_int = unsigned int; T_int shift = ((reinterpret_cast<size_t>(addr) & 3) * 8); T_int* address_uint32 = reinterpret_cast<T_int*>(addr - (reinterpret_cast<size_t>(addr) & 3)); // the 'target_value' in `old` can be different from `compare` // because other thread may update the value // before fetching a value from `address_uint32` in this function T_int old = *address_uint32; T_int assumed; T target_value; uint8_t u_val = type_reinterpret<uint8_t, T>(update_value); do { assumed = old; target_value = T((old >> shift) & 0xff); // have to compare `target_value` and `compare` before calling atomicCAS // the `target_value` in `old` can be different with `compare` if (target_value != compare) break; T_int new_value = (old & ~(0x000000ff << shift)) | (T_int(u_val) << shift); old = atomicCAS(address_uint32, assumed, new_value); } while (assumed != old); return target_value; } }; template <typename T> struct typesAtomicCASImpl<T, 2> { __forceinline__ __device__ T operator()(T* addr, T const& compare, T const& update_value) { using T_int = unsigned int; bool is_32_align = (reinterpret_cast<size_t>(addr) & 2) ? false : true; T_int* address_uint32 = reinterpret_cast<T_int*>(reinterpret_cast<size_t>(addr) - (is_32_align ? 0 : 2)); T_int old = *address_uint32; T_int assumed; T target_value; uint16_t u_val = type_reinterpret<uint16_t, T>(update_value); do { assumed = old; target_value = (is_32_align) ? T(old & 0xffff) : T(old >> 16); if (target_value != compare) break; T_int new_value = (is_32_align) ? (old & 0xffff0000) | u_val : (old & 0xffff) | (T_int(u_val) << 16); old = atomicCAS(address_uint32, assumed, new_value); } while (assumed != old); return target_value; } }; template <typename T> struct typesAtomicCASImpl<T, 4> { __forceinline__ __device__ T operator()(T* addr, T const& compare, T const& update_value) { using T_int = unsigned int; T_int ret = atomicCAS(reinterpret_cast<T_int*>(addr), type_reinterpret<T_int, T>(compare), type_reinterpret<T_int, T>(update_value)); return type_reinterpret<T, T_int>(ret); } }; // 8 bytes atomic operation template <typename T> struct typesAtomicCASImpl<T, 8> { __forceinline__ __device__ T operator()(T* addr, T const& compare, T const& update_value) { using T_int = unsigned long long int; static_assert(sizeof(T) == sizeof(T_int), errmsg_cast); T_int ret = atomicCAS(reinterpret_cast<T_int*>(addr), type_reinterpret<T_int, T>(compare), type_reinterpret<T_int, T>(update_value)); return type_reinterpret<T, T_int>(ret); } }; } // namespace detail } // namespace device_atomics /** -------------------------------------------------------------------------* * @brief compute atomic binary operation * reads the `old` located at the `address` in global or shared memory, * computes 'BinaryOp'('old', 'update_value'), * and stores the result back to memory at the same address. * These three operations are performed in one atomic transaction. * * The supported cudf types for `genericAtomicOperation` are: * int8_t, int16_t, int32_t, int64_t, float, double * * @param[in] address The address of old value in global or shared memory * @param[in] update_value The value to be computed * @param[in] op The binary operator used for compute * * @returns The old value at `address` * -------------------------------------------------------------------------**/ template <typename T, typename BinaryOp> typename std::enable_if_t<std::is_arithmetic<T>::value, T> __forceinline__ __device__ genericAtomicOperation(T* address, T const& update_value, BinaryOp op) { auto fun = raft::device_atomics::detail::genericAtomicOperationImpl<T, BinaryOp>{}; return T(fun(address, update_value, op)); } // specialization for bool types template <typename BinaryOp> __forceinline__ __device__ bool genericAtomicOperation(bool* address, bool const& update_value, BinaryOp op) { using T = bool; // don't use underlying type to apply operation for bool auto fun = raft::device_atomics::detail::genericAtomicOperationImpl<T, BinaryOp>{}; return T(fun(address, update_value, op)); } } // namespace raft /** * @brief Overloads for `atomicAdd` * * reads the `old` located at the `address` in global or shared memory, computes (old + val), and * stores the result back to memory at the same address. These three operations are performed in one * atomic transaction. * * The supported types for `atomicAdd` are: integers are floating point numbers. * CUDA natively supports `int`, `unsigned int`, `unsigned long long int`, `float`, `double. * * @param[in] address The address of old value in global or shared memory * @param[in] val The value to be added * * @returns The old value at `address` */ template <typename T> __forceinline__ __device__ T atomicAdd(T* address, T val) { return raft::genericAtomicOperation(address, val, raft::device_atomics::detail::DeviceSum{}); } /** * @brief Overloads for `atomicMin` * * reads the `old` located at the `address` in global or shared memory, computes the minimum of old * and val, and stores the result back to memory at the same address. These three operations are * performed in one atomic transaction. * * The supported types for `atomicMin` are: integers are floating point numbers. * CUDA natively supports `int`, `unsigned int`, `unsigned long long int`. * * @param[in] address The address of old value in global or shared memory * @param[in] val The value to be computed * * @returns The old value at `address` */ template <typename T> __forceinline__ __device__ T atomicMin(T* address, T val) { return raft::genericAtomicOperation(address, val, raft::device_atomics::detail::DeviceMin{}); } /** * @brief Overloads for `atomicMax` * * reads the `old` located at the `address` in global or shared memory, computes the maximum of old * and val, and stores the result back to memory at the same address. These three operations are * performed in one atomic transaction. * * The supported types for `atomicMax` are: integers are floating point numbers. * CUDA natively supports `int`, `unsigned int`, `unsigned long long int`. * * @param[in] address The address of old value in global or shared memory * @param[in] val The value to be computed * * @returns The old value at `address` */ template <typename T> __forceinline__ __device__ T atomicMax(T* address, T val) { return raft::genericAtomicOperation(address, val, raft::device_atomics::detail::DeviceMax{}); } /** * @brief Overloads for `atomicCAS` * * reads the `old` located at the `address` in global or shared memory, computes * (`old` == `compare` ? `val` : `old`), and stores the result back to memory at the same address. * These three operations are performed in one atomic transaction. * * The supported types for `atomicCAS` are: integers are floating point numbers. * CUDA natively supports `int`, `unsigned int`, `unsigned long long int`, `unsigned short int`. * * @param[in] address The address of old value in global or shared memory * @param[in] compare The value to be compared * @param[in] val The value to be computed * * @returns The old value at `address` */ template <typename T> __forceinline__ __device__ T atomicCAS(T* address, T compare, T val) { return raft::device_atomics::detail::typesAtomicCASImpl<T>()(address, compare, val); } /** * @brief Overloads for `atomicAnd` * * reads the `old` located at the `address` in global or shared memory, computes (old & val), and * stores the result back to memory at the same address. These three operations are performed in * one atomic transaction. * * The supported types for `atomicAnd` are: integers. * CUDA natively supports `int`, `unsigned int`, `unsigned long long int`. * * @param[in] address The address of old value in global or shared memory * @param[in] val The value to be computed * * @returns The old value at `address` */ template <typename T, typename std::enable_if_t<std::is_integral<T>::value, T>* = nullptr> __forceinline__ __device__ T atomicAnd(T* address, T val) { return raft::genericAtomicOperation(address, val, raft::device_atomics::detail::DeviceAnd{}); } /** * @brief Overloads for `atomicOr` * * reads the `old` located at the `address` in global or shared memory, computes (old | val), and * stores the result back to memory at the same address. These three operations are performed in * one atomic transaction. * * The supported types for `atomicOr` are: integers. * CUDA natively supports `int`, `unsigned int`, `unsigned long long int`. * * @param[in] address The address of old value in global or shared memory * @param[in] val The value to be computed * * @returns The old value at `address` */ template <typename T, typename std::enable_if_t<std::is_integral<T>::value, T>* = nullptr> __forceinline__ __device__ T atomicOr(T* address, T val) { return raft::genericAtomicOperation(address, val, raft::device_atomics::detail::DeviceOr{}); } /** * @brief Overloads for `atomicXor` * * reads the `old` located at the `address` in global or shared memory, computes (old ^ val), and * stores the result back to memory at the same address. These three operations are performed in * one atomic transaction. * * The supported types for `atomicXor` are: integers. * CUDA natively supports `int`, `unsigned int`, `unsigned long long int`. * * @param[in] address The address of old value in global or shared memory * @param[in] val The value to be computed * * @returns The old value at `address` */ template <typename T, typename std::enable_if_t<std::is_integral<T>::value, T>* = nullptr> __forceinline__ __device__ T atomicXor(T* address, T val) { return raft::genericAtomicOperation(address, val, raft::device_atomics::detail::DeviceXor{}); } /** * @brief: Warp aggregated atomic increment * * increments an atomic counter using all active threads in a warp. The return * value is the original value of the counter plus the rank of the calling * thread. * * The use of atomicIncWarp is a performance optimization. It can reduce the * amount of atomic memory traffic by a factor of 32. * * Adapted from: * https://developer.nvidia.com/blog/cuda-pro-tip-optimized-filtering-warp-aggregated-atomics/ * * @tparam T An integral type * @param[in,out] ctr The address of old value * * @return The old value of the counter plus the rank of the calling thread. */ template <typename T = unsigned int, typename std::enable_if_t<std::is_integral<T>::value, T>* = nullptr> __device__ T atomicIncWarp(T* ctr) { namespace cg = cooperative_groups; auto g = cg::coalesced_threads(); T warp_res; if (g.thread_rank() == 0) { warp_res = atomicAdd(ctr, static_cast<T>(g.size())); } return g.shfl(warp_res, 0) + g.thread_rank(); }
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/device_loads_stores.cuh
/* * Copyright (c) 2021-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <cstdint> // uintX_t #include <raft/core/device_span.hpp> #include <raft/util/cuda_dev_essentials.cuh> // DI namespace raft { /** * @defgroup SmemStores Shared memory store operations * @{ * @brief Stores to shared memory (both vectorized and non-vectorized forms) * requires the given shmem pointer to be aligned by the vector length, like for float4 lds/sts shmem pointer should be aligned by 16 bytes else it might silently fail or can also give runtime error. * @param[out] addr shared memory address (should be aligned to vector size) * @param[in] x data to be stored at this address */ DI void sts(uint8_t* addr, const uint8_t& x) { uint32_t x_int; x_int = x; auto s1 = __cvta_generic_to_shared(reinterpret_cast<uint8_t*>(addr)); asm volatile("st.shared.u8 [%0], {%1};" : : "l"(s1), "r"(x_int)); } DI void sts(uint8_t* addr, const uint8_t (&x)[1]) { uint32_t x_int[1]; x_int[0] = x[0]; auto s1 = __cvta_generic_to_shared(reinterpret_cast<uint8_t*>(addr)); asm volatile("st.shared.u8 [%0], {%1};" : : "l"(s1), "r"(x_int[0])); } DI void sts(uint8_t* addr, const uint8_t (&x)[2]) { uint32_t x_int[2]; x_int[0] = x[0]; x_int[1] = x[1]; auto s2 = __cvta_generic_to_shared(reinterpret_cast<uint8_t*>(addr)); asm volatile("st.shared.v2.u8 [%0], {%1, %2};" : : "l"(s2), "r"(x_int[0]), "r"(x_int[1])); } DI void sts(uint8_t* addr, const uint8_t (&x)[4]) { uint32_t x_int[4]; x_int[0] = x[0]; x_int[1] = x[1]; x_int[2] = x[2]; x_int[3] = x[3]; auto s4 = __cvta_generic_to_shared(reinterpret_cast<uint8_t*>(addr)); asm volatile("st.shared.v4.u8 [%0], {%1, %2, %3, %4};" : : "l"(s4), "r"(x_int[0]), "r"(x_int[1]), "r"(x_int[2]), "r"(x_int[3])); } DI void sts(int8_t* addr, const int8_t& x) { int32_t x_int = x; auto s1 = __cvta_generic_to_shared(reinterpret_cast<int8_t*>(addr)); asm volatile("st.shared.s8 [%0], {%1};" : : "l"(s1), "r"(x_int)); } DI void sts(int8_t* addr, const int8_t (&x)[1]) { int32_t x_int[1]; x_int[0] = x[0]; auto s1 = __cvta_generic_to_shared(reinterpret_cast<int8_t*>(addr)); asm volatile("st.shared.s8 [%0], {%1};" : : "l"(s1), "r"(x_int[0])); } DI void sts(int8_t* addr, const int8_t (&x)[2]) { int32_t x_int[2]; x_int[0] = x[0]; x_int[1] = x[1]; auto s2 = __cvta_generic_to_shared(reinterpret_cast<int8_t*>(addr)); asm volatile("st.shared.v2.s8 [%0], {%1, %2};" : : "l"(s2), "r"(x_int[0]), "r"(x_int[1])); } DI void sts(int8_t* addr, const int8_t (&x)[4]) { int32_t x_int[4]; x_int[0] = x[0]; x_int[1] = x[1]; x_int[2] = x[2]; x_int[3] = x[3]; auto s4 = __cvta_generic_to_shared(reinterpret_cast<int8_t*>(addr)); asm volatile("st.shared.v4.s8 [%0], {%1, %2, %3, %4};" : : "l"(s4), "r"(x_int[0]), "r"(x_int[1]), "r"(x_int[2]), "r"(x_int[3])); } DI void sts(uint32_t* addr, const uint32_t& x) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<uint32_t*>(addr)); asm volatile("st.shared.u32 [%0], {%1};" : : "l"(s1), "r"(x)); } DI void sts(uint32_t* addr, const uint32_t (&x)[1]) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<uint32_t*>(addr)); asm volatile("st.shared.u32 [%0], {%1};" : : "l"(s1), "r"(x[0])); } DI void sts(uint32_t* addr, const uint32_t (&x)[2]) { auto s2 = __cvta_generic_to_shared(reinterpret_cast<uint2*>(addr)); asm volatile("st.shared.v2.u32 [%0], {%1, %2};" : : "l"(s2), "r"(x[0]), "r"(x[1])); } DI void sts(uint32_t* addr, const uint32_t (&x)[4]) { auto s4 = __cvta_generic_to_shared(reinterpret_cast<uint4*>(addr)); asm volatile("st.shared.v4.u32 [%0], {%1, %2, %3, %4};" : : "l"(s4), "r"(x[0]), "r"(x[1]), "r"(x[2]), "r"(x[3])); } DI void sts(int32_t* addr, const int32_t& x) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<int32_t*>(addr)); asm volatile("st.shared.u32 [%0], {%1};" : : "l"(s1), "r"(x)); } DI void sts(int32_t* addr, const int32_t (&x)[1]) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<int32_t*>(addr)); asm volatile("st.shared.u32 [%0], {%1};" : : "l"(s1), "r"(x[0])); } DI void sts(int32_t* addr, const int32_t (&x)[2]) { auto s2 = __cvta_generic_to_shared(reinterpret_cast<int2*>(addr)); asm volatile("st.shared.v2.u32 [%0], {%1, %2};" : : "l"(s2), "r"(x[0]), "r"(x[1])); } DI void sts(int32_t* addr, const int32_t (&x)[4]) { auto s4 = __cvta_generic_to_shared(reinterpret_cast<int4*>(addr)); asm volatile("st.shared.v4.u32 [%0], {%1, %2, %3, %4};" : : "l"(s4), "r"(x[0]), "r"(x[1]), "r"(x[2]), "r"(x[3])); } DI void sts(float* addr, const float& x) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<float*>(addr)); asm volatile("st.shared.f32 [%0], {%1};" : : "l"(s1), "f"(x)); } DI void sts(float* addr, const float (&x)[1]) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<float*>(addr)); asm volatile("st.shared.f32 [%0], {%1};" : : "l"(s1), "f"(x[0])); } DI void sts(float* addr, const float (&x)[2]) { auto s2 = __cvta_generic_to_shared(reinterpret_cast<float2*>(addr)); asm volatile("st.shared.v2.f32 [%0], {%1, %2};" : : "l"(s2), "f"(x[0]), "f"(x[1])); } DI void sts(float* addr, const float (&x)[4]) { auto s4 = __cvta_generic_to_shared(reinterpret_cast<float4*>(addr)); asm volatile("st.shared.v4.f32 [%0], {%1, %2, %3, %4};" : : "l"(s4), "f"(x[0]), "f"(x[1]), "f"(x[2]), "f"(x[3])); } DI void sts(double* addr, const double& x) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<double*>(addr)); asm volatile("st.shared.f64 [%0], {%1};" : : "l"(s1), "d"(x)); } DI void sts(double* addr, const double (&x)[1]) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<double*>(addr)); asm volatile("st.shared.f64 [%0], {%1};" : : "l"(s1), "d"(x[0])); } DI void sts(double* addr, const double (&x)[2]) { auto s2 = __cvta_generic_to_shared(reinterpret_cast<double2*>(addr)); asm volatile("st.shared.v2.f64 [%0], {%1, %2};" : : "l"(s2), "d"(x[0]), "d"(x[1])); } /** @} */ /** * @defgroup SmemLoads Shared memory load operations * @{ * @brief Loads from shared memory (both vectorized and non-vectorized forms) requires the given shmem pointer to be aligned by the vector length, like for float4 lds/sts shmem pointer should be aligned by 16 bytes else it might silently fail or can also give runtime error. * @param[out] x the data to be loaded * @param[in] addr shared memory address from where to load * (should be aligned to vector size) */ DI void lds(uint8_t& x, const uint8_t* addr) { uint32_t x_int; auto s1 = __cvta_generic_to_shared(reinterpret_cast<const uint8_t*>(addr)); asm volatile("ld.shared.u8 {%0}, [%1];" : "=r"(x_int) : "l"(s1)); x = x_int; } DI void lds(uint8_t (&x)[1], const uint8_t* addr) { uint32_t x_int[1]; auto s1 = __cvta_generic_to_shared(reinterpret_cast<const uint8_t*>(addr)); asm volatile("ld.shared.u8 {%0}, [%1];" : "=r"(x_int[0]) : "l"(s1)); x[0] = x_int[0]; } DI void lds(uint8_t (&x)[2], const uint8_t* addr) { uint32_t x_int[2]; auto s2 = __cvta_generic_to_shared(reinterpret_cast<const uint8_t*>(addr)); asm volatile("ld.shared.v2.u8 {%0, %1}, [%2];" : "=r"(x_int[0]), "=r"(x_int[1]) : "l"(s2)); x[0] = x_int[0]; x[1] = x_int[1]; } DI void lds(uint8_t (&x)[4], const uint8_t* addr) { uint32_t x_int[4]; auto s4 = __cvta_generic_to_shared(reinterpret_cast<const uint8_t*>(addr)); asm volatile("ld.shared.v4.u8 {%0, %1, %2, %3}, [%4];" : "=r"(x_int[0]), "=r"(x_int[1]), "=r"(x_int[2]), "=r"(x_int[3]) : "l"(s4)); x[0] = x_int[0]; x[1] = x_int[1]; x[2] = x_int[2]; x[3] = x_int[3]; } DI void lds(int8_t& x, const int8_t* addr) { int32_t x_int; auto s1 = __cvta_generic_to_shared(reinterpret_cast<const int8_t*>(addr)); asm volatile("ld.shared.s8 {%0}, [%1];" : "=r"(x_int) : "l"(s1)); x = x_int; } DI void lds(int8_t (&x)[1], const int8_t* addr) { int32_t x_int[1]; auto s1 = __cvta_generic_to_shared(reinterpret_cast<const int8_t*>(addr)); asm volatile("ld.shared.s8 {%0}, [%1];" : "=r"(x_int[0]) : "l"(s1)); x[0] = x_int[0]; } DI void lds(int8_t (&x)[2], const int8_t* addr) { int32_t x_int[2]; auto s2 = __cvta_generic_to_shared(reinterpret_cast<const int8_t*>(addr)); asm volatile("ld.shared.v2.s8 {%0, %1}, [%2];" : "=r"(x_int[0]), "=r"(x_int[1]) : "l"(s2)); x[0] = x_int[0]; x[1] = x_int[1]; } DI void lds(int8_t (&x)[4], const int8_t* addr) { int32_t x_int[4]; auto s4 = __cvta_generic_to_shared(reinterpret_cast<const int8_t*>(addr)); asm volatile("ld.shared.v4.s8 {%0, %1, %2, %3}, [%4];" : "=r"(x_int[0]), "=r"(x_int[1]), "=r"(x_int[2]), "=r"(x_int[3]) : "l"(s4)); x[0] = x_int[0]; x[1] = x_int[1]; x[2] = x_int[2]; x[3] = x_int[3]; } DI void lds(uint32_t (&x)[4], const uint32_t* addr) { auto s4 = __cvta_generic_to_shared(reinterpret_cast<const uint32_t*>(addr)); asm volatile("ld.shared.v4.u32 {%0, %1, %2, %3}, [%4];" : "=r"(x[0]), "=r"(x[1]), "=r"(x[2]), "=r"(x[3]) : "l"(s4)); } DI void lds(uint32_t (&x)[2], const uint32_t* addr) { auto s2 = __cvta_generic_to_shared(reinterpret_cast<const uint32_t*>(addr)); asm volatile("ld.shared.v2.u32 {%0, %1}, [%2];" : "=r"(x[0]), "=r"(x[1]) : "l"(s2)); } DI void lds(uint32_t (&x)[1], const uint32_t* addr) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<const uint32_t*>(addr)); asm volatile("ld.shared.u32 {%0}, [%1];" : "=r"(x[0]) : "l"(s1)); } DI void lds(uint32_t& x, const uint32_t* addr) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<const uint32_t*>(addr)); asm volatile("ld.shared.u32 {%0}, [%1];" : "=r"(x) : "l"(s1)); } DI void lds(int32_t (&x)[4], const int32_t* addr) { auto s4 = __cvta_generic_to_shared(reinterpret_cast<const int32_t*>(addr)); asm volatile("ld.shared.v4.u32 {%0, %1, %2, %3}, [%4];" : "=r"(x[0]), "=r"(x[1]), "=r"(x[2]), "=r"(x[3]) : "l"(s4)); } DI void lds(int32_t (&x)[2], const int32_t* addr) { auto s2 = __cvta_generic_to_shared(reinterpret_cast<const int32_t*>(addr)); asm volatile("ld.shared.v2.u32 {%0, %1}, [%2];" : "=r"(x[0]), "=r"(x[1]) : "l"(s2)); } DI void lds(int32_t (&x)[1], const int32_t* addr) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<const int32_t*>(addr)); asm volatile("ld.shared.u32 {%0}, [%1];" : "=r"(x[0]) : "l"(s1)); } DI void lds(int32_t& x, const int32_t* addr) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<const int32_t*>(addr)); asm volatile("ld.shared.u32 {%0}, [%1];" : "=r"(x) : "l"(s1)); } DI void lds(float& x, const float* addr) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<const float*>(addr)); asm volatile("ld.shared.f32 {%0}, [%1];" : "=f"(x) : "l"(s1)); } DI void lds(float (&x)[1], const float* addr) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<const float*>(addr)); asm volatile("ld.shared.f32 {%0}, [%1];" : "=f"(x[0]) : "l"(s1)); } DI void lds(float (&x)[2], const float* addr) { auto s2 = __cvta_generic_to_shared(reinterpret_cast<const float2*>(addr)); asm volatile("ld.shared.v2.f32 {%0, %1}, [%2];" : "=f"(x[0]), "=f"(x[1]) : "l"(s2)); } DI void lds(float (&x)[4], const float* addr) { auto s4 = __cvta_generic_to_shared(reinterpret_cast<const float4*>(addr)); asm volatile("ld.shared.v4.f32 {%0, %1, %2, %3}, [%4];" : "=f"(x[0]), "=f"(x[1]), "=f"(x[2]), "=f"(x[3]) : "l"(s4)); } DI void lds(float& x, float* addr) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<float*>(addr)); asm volatile("ld.shared.f32 {%0}, [%1];" : "=f"(x) : "l"(s1)); } DI void lds(float (&x)[1], float* addr) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<float*>(addr)); asm volatile("ld.shared.f32 {%0}, [%1];" : "=f"(x[0]) : "l"(s1)); } DI void lds(float (&x)[2], float* addr) { auto s2 = __cvta_generic_to_shared(reinterpret_cast<float2*>(addr)); asm volatile("ld.shared.v2.f32 {%0, %1}, [%2];" : "=f"(x[0]), "=f"(x[1]) : "l"(s2)); } DI void lds(float (&x)[4], float* addr) { auto s4 = __cvta_generic_to_shared(reinterpret_cast<float4*>(addr)); asm volatile("ld.shared.v4.f32 {%0, %1, %2, %3}, [%4];" : "=f"(x[0]), "=f"(x[1]), "=f"(x[2]), "=f"(x[3]) : "l"(s4)); } DI void lds(double& x, double* addr) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<double*>(addr)); asm volatile("ld.shared.f64 {%0}, [%1];" : "=d"(x) : "l"(s1)); } DI void lds(double (&x)[1], double* addr) { auto s1 = __cvta_generic_to_shared(reinterpret_cast<double*>(addr)); asm volatile("ld.shared.f64 {%0}, [%1];" : "=d"(x[0]) : "l"(s1)); } DI void lds(double (&x)[2], double* addr) { auto s2 = __cvta_generic_to_shared(reinterpret_cast<double2*>(addr)); asm volatile("ld.shared.v2.f64 {%0, %1}, [%2];" : "=d"(x[0]), "=d"(x[1]) : "l"(s2)); } /** @} */ /** * @defgroup GlobalLoads Global cached load operations * @{ * @brief Load from global memory with caching at L1 level * @param[out] x data to be loaded from global memory * @param[in] addr address in global memory from where to load */ DI void ldg(float& x, const float* addr) { asm volatile("ld.global.cg.f32 %0, [%1];" : "=f"(x) : "l"(addr)); } DI void ldg(float (&x)[1], const float* addr) { asm volatile("ld.global.cg.f32 %0, [%1];" : "=f"(x[0]) : "l"(addr)); } DI void ldg(float (&x)[2], const float* addr) { asm volatile("ld.global.cg.v2.f32 {%0, %1}, [%2];" : "=f"(x[0]), "=f"(x[1]) : "l"(addr)); } DI void ldg(float (&x)[4], const float* addr) { asm volatile("ld.global.cg.v4.f32 {%0, %1, %2, %3}, [%4];" : "=f"(x[0]), "=f"(x[1]), "=f"(x[2]), "=f"(x[3]) : "l"(addr)); } DI void ldg(double& x, const double* addr) { asm volatile("ld.global.cg.f64 %0, [%1];" : "=d"(x) : "l"(addr)); } DI void ldg(double (&x)[1], const double* addr) { asm volatile("ld.global.cg.f64 %0, [%1];" : "=d"(x[0]) : "l"(addr)); } DI void ldg(double (&x)[2], const double* addr) { asm volatile("ld.global.cg.v2.f64 {%0, %1}, [%2];" : "=d"(x[0]), "=d"(x[1]) : "l"(addr)); } DI void ldg(uint32_t (&x)[4], const uint32_t* const& addr) { asm volatile("ld.global.cg.v4.u32 {%0, %1, %2, %3}, [%4];" : "=r"(x[0]), "=r"(x[1]), "=r"(x[2]), "=r"(x[3]) : "l"(addr)); } DI void ldg(uint32_t (&x)[2], const uint32_t* const& addr) { asm volatile("ld.global.cg.v2.u32 {%0, %1}, [%2];" : "=r"(x[0]), "=r"(x[1]) : "l"(addr)); } DI void ldg(uint32_t (&x)[1], const uint32_t* const& addr) { asm volatile("ld.global.cg.u32 %0, [%1];" : "=r"(x[0]) : "l"(addr)); } DI void ldg(uint32_t& x, const uint32_t* const& addr) { asm volatile("ld.global.cg.u32 %0, [%1];" : "=r"(x) : "l"(addr)); } DI void ldg(int32_t (&x)[4], const int32_t* const& addr) { asm volatile("ld.global.cg.v4.u32 {%0, %1, %2, %3}, [%4];" : "=r"(x[0]), "=r"(x[1]), "=r"(x[2]), "=r"(x[3]) : "l"(addr)); } DI void ldg(int32_t (&x)[2], const int32_t* const& addr) { asm volatile("ld.global.cg.v2.u32 {%0, %1}, [%2];" : "=r"(x[0]), "=r"(x[1]) : "l"(addr)); } DI void ldg(int32_t (&x)[1], const int32_t* const& addr) { asm volatile("ld.global.cg.u32 %0, [%1];" : "=r"(x[0]) : "l"(addr)); } DI void ldg(int32_t& x, const int32_t* const& addr) { asm volatile("ld.global.cg.u32 %0, [%1];" : "=r"(x) : "l"(addr)); } DI void ldg(uint8_t (&x)[4], const uint8_t* const& addr) { uint32_t x_int[4]; asm volatile("ld.global.cg.v4.u8 {%0, %1, %2, %3}, [%4];" : "=r"(x_int[0]), "=r"(x_int[1]), "=r"(x_int[2]), "=r"(x_int[3]) : "l"(addr)); x[0] = x_int[0]; x[1] = x_int[1]; x[2] = x_int[2]; x[3] = x_int[3]; } DI void ldg(uint8_t (&x)[2], const uint8_t* const& addr) { uint32_t x_int[2]; asm volatile("ld.global.cg.v2.u8 {%0, %1}, [%2];" : "=r"(x_int[0]), "=r"(x_int[1]) : "l"(addr)); x[0] = x_int[0]; x[1] = x_int[1]; } DI void ldg(uint8_t (&x)[1], const uint8_t* const& addr) { uint32_t x_int; asm volatile("ld.global.cg.u8 %0, [%1];" : "=r"(x_int) : "l"(addr)); x[0] = x_int; } DI void ldg(uint8_t& x, const uint8_t* const& addr) { uint32_t x_int; asm volatile("ld.global.cg.u8 %0, [%1];" : "=r"(x_int) : "l"(addr)); x = x_int; } DI void ldg(int8_t (&x)[4], const int8_t* const& addr) { int x_int[4]; asm volatile("ld.global.cg.v4.s8 {%0, %1, %2, %3}, [%4];" : "=r"(x_int[0]), "=r"(x_int[1]), "=r"(x_int[2]), "=r"(x_int[3]) : "l"(addr)); x[0] = x_int[0]; x[1] = x_int[1]; x[2] = x_int[2]; x[3] = x_int[3]; } DI void ldg(int8_t (&x)[2], const int8_t* const& addr) { int x_int[2]; asm volatile("ld.global.cg.v2.s8 {%0, %1}, [%2];" : "=r"(x_int[0]), "=r"(x_int[1]) : "l"(addr)); x[0] = x_int[0]; x[1] = x_int[1]; } DI void ldg(int8_t& x, const int8_t* const& addr) { int x_int; asm volatile("ld.global.cg.s8 %0, [%1];" : "=r"(x_int) : "l"(addr)); x = x_int; } DI void ldg(int8_t (&x)[1], const int8_t* const& addr) { int x_int; asm volatile("ld.global.cg.s8 %0, [%1];" : "=r"(x_int) : "l"(addr)); x[0] = x_int; } /** * @brief Executes a 1D block strided copy * @param dst destination pointer * @param src source pointer * @param size number of items to copy */ template <typename T> DI void block_copy(T* dst, const T* src, const size_t size) { for (auto i = threadIdx.x; i < size; i += blockDim.x) { dst[i] = src[i]; } } /** * @brief Executes a 1D block strided copy * @param dst span of destination pointer * @param src span of source pointer * @param size number of items to copy */ template <typename T> DI void block_copy(raft::device_span<T> dst, const raft::device_span<const T> src, const size_t size) { assert(src.size() >= size); assert(dst.size() >= size); block_copy(dst.data(), src.data(), size); } /** * @brief Executes a 1D block strided copy * @param dst span of destination pointer * @param src span of source pointer * @param size number of items to copy */ template <typename T> DI void block_copy(raft::device_span<T> dst, const raft::device_span<T> src, const size_t size) { assert(src.size() >= size); assert(dst.size() >= size); block_copy(dst.data(), src.data(), size); } /** * @brief Executes a 1D block strided copy * @param dst span of destination pointer * @param src span of source pointer */ template <typename T> DI void block_copy(raft::device_span<T> dst, const raft::device_span<T> src) { assert(dst.size() >= src.size()); block_copy(dst, src, src.size()); } /** @} */ } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/device_utils.cuh
/* * Copyright (c) 2021-2022, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/util/cuda_utils.cuh> #include <utility> // pair namespace raft { // TODO move to raft https://github.com/rapidsai/raft/issues/90 /** helper method to get the compute capability version numbers */ inline std::pair<int, int> getDeviceCapability() { int devId; RAFT_CUDA_TRY(cudaGetDevice(&devId)); int major, minor; RAFT_CUDA_TRY(cudaDeviceGetAttribute(&major, cudaDevAttrComputeCapabilityMajor, devId)); RAFT_CUDA_TRY(cudaDeviceGetAttribute(&minor, cudaDevAttrComputeCapabilityMinor, devId)); return std::make_pair(major, minor); } /** * @brief Batched warp-level sum reduction * * @tparam T data type * @tparam NThreads Number of threads in the warp doing independent reductions * * @param[in] val input value * @return for the first "group" of threads, the reduced value. All * others will contain unusable values! * * @note Why not cub? Because cub doesn't seem to allow working with arbitrary * number of warps in a block and also doesn't support this kind of * batched reduction operation * @note All threads in the warp must enter this function together * * @todo Expand this to support arbitrary reduction ops */ template <typename T, int NThreads> DI T batchedWarpReduce(T val) { #pragma unroll for (int i = NThreads; i < raft::WarpSize; i <<= 1) { val += raft::shfl(val, raft::laneId() + i); } return val; } /** * @brief 1-D block-level batched sum reduction * * @tparam T data type * @tparam NThreads Number of threads in the warp doing independent reductions * * @param val input value * @param smem shared memory region needed for storing intermediate results. It * must alteast be of size: `sizeof(T) * nWarps * NThreads` * @return for the first "group" of threads in the block, the reduced value. * All others will contain unusable values! * * @note Why not cub? Because cub doesn't seem to allow working with arbitrary * number of warps in a block and also doesn't support this kind of * batched reduction operation * @note All threads in the block must enter this function together * * @todo Expand this to support arbitrary reduction ops */ template <typename T, int NThreads> DI T batchedBlockReduce(T val, char* smem) { auto* sTemp = reinterpret_cast<T*>(smem); constexpr int nGroupsPerWarp = raft::WarpSize / NThreads; static_assert(raft::isPo2(nGroupsPerWarp), "nGroupsPerWarp must be a PO2!"); const int nGroups = (blockDim.x + NThreads - 1) / NThreads; const int lid = raft::laneId(); const int lgid = lid % NThreads; const int gid = threadIdx.x / NThreads; const auto wrIdx = (gid / nGroupsPerWarp) * NThreads + lgid; const auto rdIdx = gid * NThreads + lgid; for (int i = nGroups; i > 0;) { auto iAligned = ((i + nGroupsPerWarp - 1) / nGroupsPerWarp) * nGroupsPerWarp; if (gid < iAligned) { val = batchedWarpReduce<T, NThreads>(val); if (lid < NThreads) sTemp[wrIdx] = val; } __syncthreads(); i /= nGroupsPerWarp; if (i > 0) { val = gid < i ? sTemp[rdIdx] : T(0); } __syncthreads(); } return val; } } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/cutlass_utils.cuh
/* * Copyright (c) 2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <cutlass/cutlass.h> #include <raft/core/error.hpp> namespace raft { /** * @brief Exception thrown when a CUTLASS error is encountered. */ struct cutlass_error : public raft::exception { explicit cutlass_error(char const* const message) : raft::exception(message) {} explicit cutlass_error(std::string const& message) : raft::exception(message) {} }; } // namespace raft /** * @brief Error checking macro for CUTLASS functions. * * Invokes a CUTLASS function call, if the call does not return cutlass::Status::kSuccess, * throws an exception detailing the CUTLASS error that occurred. * */ #define RAFT_CUTLASS_TRY(call) \ do { \ cutlass::Status const status = call; \ if (status != cutlass::Status::kSuccess) { \ std::string msg{}; \ SET_ERROR_MSG(msg, \ "CUTLASS error encountered at: ", \ "call='%s', Reason=%s", \ #call, \ cutlassGetStatusString(status)); \ throw raft::cutlass_error(msg); \ } \ } while (0)
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/bitonic_sort.cuh
/* * Copyright (c) 2022-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/core/detail/macros.hpp> #include <raft/util/cuda_utils.cuh> namespace raft::util { namespace { template <typename T> _RAFT_DEVICE _RAFT_FORCEINLINE void swap(T& x, T& y) { T t = x; x = y; y = t; } template <typename T> _RAFT_DEVICE _RAFT_FORCEINLINE void conditional_assign(bool cond, T& ptr, T x) { if (cond) { ptr = x; } } } // namespace /** * Warp-wide bitonic merge and sort. * The data is strided among `warp_width` threads, * e.g. calling `bitonic<4>(ascending=true).sort(arr)` takes a unique 4-element array as input of * each thread in a warp and sorts them, such that for a fixed i, arr[i] are sorted within the * threads in a warp, and for any i < j, arr[j] in any thread is not smaller than arr[i] in any * other thread. * When `warp_width < WarpSize`, the data is sorted within all subwarps of the warp independently. * * As an example, assuming `Size = 4`, `warp_width = 16`, and `WarpSize = 32`, sorting a permutation * of numbers 0-63 in each subwarp yield the following result: * ` * arr_i \ laneId() * 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 ... * subwarp_1 subwarp_2 * 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 ... * 1 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 16 17 18 ... * 2 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 32 33 34 ... * 3 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 48 49 50 ... * ` * * Here is a small usage example of device code, which sorts the arrays of length 8 (= 4 * 2) * grouped in pairs of threads in ascending order: * @code{.cpp} * // Fill an array of four ints in each thread of a warp. * int i = laneId(); * int arr[4] = {i+1, i+5, i, i+7}; * // Sort the arrays in groups of two threads. * bitonic<4>(ascending=true, warp_width=2).sort(arr); * // As a result, * // for every even thread (`i == 2j`): arr == {2j, 2j+1, 2j+5, 2j+7} * // for every odd thread (`i == 2j+1`): arr == {2j+1, 2j+2, 2j+6, 2j+8} * @endcode * * @tparam Size * number of elements processed in each thread; * i.e. the total data size is `Size * warp_width`. * Must be power-of-two. * */ template <int Size = 1> class bitonic { static_assert(isPo2(Size)); public: /** * Initialize bitonic sort config. * * @param ascending * the resulting order (true: ascending, false: descending). * @param warp_width * the number of threads participating in the warp-level primitives; * the total size of the sorted data is `Size * warp_width`. * Must be power-of-two, not larger than the WarpSize. */ _RAFT_DEVICE _RAFT_FORCEINLINE explicit bitonic(bool ascending, int warp_width = WarpSize) : ascending_(ascending), warp_width_(warp_width) { } bitonic(bitonic const&) = delete; bitonic(bitonic&&) = delete; auto operator=(bitonic const&) -> bitonic& = delete; auto operator=(bitonic&&) -> bitonic& = delete; /** * You can think of this function in two ways: * * 1) Sort any bitonic sequence. * 2) Merge two halves of the input data assuming they're already sorted, and their order is * opposite (i.e. either ascending+descending or descending+ascending). * * The input pointers are unique per-thread. * See the class description for the description of the data layout. * * @param keys * is a device pointer to a contiguous array of keys, unique per thread; must be at least `Size` * elements long. * @param payloads * are zero or more associated arrays of the same size as keys, which are sorted together with * the keys; must be at least `Size` elements long. */ template <typename KeyT, typename... PayloadTs> _RAFT_DEVICE _RAFT_FORCEINLINE void merge(KeyT* __restrict__ keys, PayloadTs* __restrict__... payloads) const { return bitonic<Size>::merge_impl(ascending_, warp_width_, keys, payloads...); } /** * Sort the data. * The input pointers are unique per-thread. * See the class description for the description of the data layout. * * @param keys * is a device pointer to a contiguous array of keys, unique per thread; must be at least `Size` * elements long. * @param payloads * are zero or more associated arrays of the same size as keys, which are sorted together with * the keys; must be at least `Size` elements long. */ template <typename KeyT, typename... PayloadTs> _RAFT_DEVICE _RAFT_FORCEINLINE void sort(KeyT* __restrict__ keys, PayloadTs* __restrict__... payloads) const { return bitonic<Size>::sort_impl(ascending_, warp_width_, keys, payloads...); } /** * @brief `merge` variant for the case of one element per thread * (pass input by a reference instead of a pointer). * * @param key * @param payload */ template <typename KeyT, typename... PayloadTs, int S = Size> _RAFT_DEVICE _RAFT_FORCEINLINE auto merge(KeyT& __restrict__ key, PayloadTs& __restrict__... payload) const -> std::enable_if_t<S == 1, void> // SFINAE to enable this for Size == 1 only { static_assert(S == Size); return merge(&key, &payload...); } /** * @brief `sort` variant for the case of one element per thread * (pass input by a reference instead of a pointer). * * @param key * @param payload */ template <typename KeyT, typename... PayloadTs, int S = Size> _RAFT_DEVICE _RAFT_FORCEINLINE auto sort(KeyT& __restrict__ key, PayloadTs& __restrict__... payload) const -> std::enable_if_t<S == 1, void> // SFINAE to enable this for Size == 1 only { static_assert(S == Size); return sort(&key, &payload...); } private: const int warp_width_; const bool ascending_; template <int AnotherSize> friend class bitonic; template <typename KeyT, typename... PayloadTs> static _RAFT_DEVICE _RAFT_FORCEINLINE void merge_impl(bool ascending, int warp_width, KeyT* __restrict__ keys, PayloadTs* __restrict__... payloads) { #pragma unroll for (int size = Size; size > 1; size >>= 1) { const int stride = size >> 1; #pragma unroll for (int offset = 0; offset < Size; offset += size) { #pragma unroll for (int i = offset + stride - 1; i >= offset; i--) { const int other_i = i + stride; KeyT& key = keys[i]; KeyT& other = keys[other_i]; if (ascending ? key > other : key < other) { swap(key, other); (swap(payloads[i], payloads[other_i]), ...); } } } } const int lane = laneId(); #pragma unroll for (int i = 0; i < Size; i++) { KeyT& key = keys[i]; for (int stride = (warp_width >> 1); stride > 0; stride >>= 1) { const bool is_second = lane & stride; const KeyT other = shfl_xor(key, stride, warp_width); const bool do_assign = (ascending != is_second) ? key > other : key < other; conditional_assign(do_assign, key, other); // NB: don't put shfl_xor in a conditional; it must be called by all threads in a warp. (conditional_assign(do_assign, payloads[i], shfl_xor(payloads[i], stride, warp_width)), ...); } } } template <typename KeyT, typename... PayloadTs> static _RAFT_DEVICE _RAFT_FORCEINLINE void sort_impl(bool ascending, int warp_width, KeyT* __restrict__ keys, PayloadTs* __restrict__... payloads) { if constexpr (Size == 1) { const int lane = laneId(); for (int width = 2; width < warp_width; width <<= 1) { bitonic<1>::merge_impl(lane & width, width, keys, payloads...); } } else { constexpr int kSize2 = Size / 2; bitonic<kSize2>::sort_impl(false, warp_width, keys, payloads...); bitonic<kSize2>::sort_impl(true, warp_width, keys + kSize2, (payloads + kSize2)...); } bitonic<Size>::merge_impl(ascending, warp_width, keys, payloads...); } }; } // namespace raft::util
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/scatter.cuh
/* * Copyright (c) 2019-2022, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/core/operators.hpp> #include <raft/util/cuda_utils.cuh> #include <raft/util/detail/scatter.cuh> namespace raft { /** * @brief Performs scatter operation based on the input indexing array * @tparam DataT data type whose array gets scattered * @tparam IdxT indexing type * @tparam TPB threads-per-block in the final kernel launched * @tparam Lambda the device-lambda performing a unary operation on the loaded * data before it gets scattered * @param out the output array * @param in the input array * @param idx the indexing array * @param len number of elements in the input array * @param stream cuda stream where to launch work * @param op the device-lambda with signature `DataT func(DataT, IdxT);`. This * will be applied to every element before scattering it to the right location. * The second param in this method will be the destination index. */ template <typename DataT, typename IdxT, typename Lambda = raft::identity_op, int TPB = 256> void scatter(DataT* out, const DataT* in, const IdxT* idx, IdxT len, cudaStream_t stream, Lambda op = raft::identity_op()) { if (len <= 0) return; constexpr size_t DataSize = sizeof(DataT); constexpr size_t IdxSize = sizeof(IdxT); constexpr size_t MaxPerElem = DataSize > IdxSize ? DataSize : IdxSize; size_t bytes = len * MaxPerElem; if (16 / MaxPerElem && bytes % 16 == 0) { detail::scatterImpl<DataT, 16 / MaxPerElem, Lambda, IdxT, TPB>(out, in, idx, len, op, stream); } else if (8 / MaxPerElem && bytes % 8 == 0) { detail::scatterImpl<DataT, 8 / MaxPerElem, Lambda, IdxT, TPB>(out, in, idx, len, op, stream); } else if (4 / MaxPerElem && bytes % 4 == 0) { detail::scatterImpl<DataT, 4 / MaxPerElem, Lambda, IdxT, TPB>(out, in, idx, len, op, stream); } else if (2 / MaxPerElem && bytes % 2 == 0) { detail::scatterImpl<DataT, 2 / MaxPerElem, Lambda, IdxT, TPB>(out, in, idx, len, op, stream); } else if (1 / MaxPerElem) { detail::scatterImpl<DataT, 1 / MaxPerElem, Lambda, IdxT, TPB>(out, in, idx, len, op, stream); } else { detail::scatterImpl<DataT, 1, Lambda, IdxT, TPB>(out, in, idx, len, op, stream); } } } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/vectorized.cuh
/* * Copyright (c) 2018-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <cuda_fp16.h> #include <raft/util/cuda_utils.cuh> namespace raft { template <typename math_, int VecLen> struct IOType {}; template <> struct IOType<bool, 1> { static_assert(sizeof(bool) == sizeof(int8_t), "IOType bool size assumption failed"); typedef int8_t Type; }; template <> struct IOType<bool, 2> { typedef int16_t Type; }; template <> struct IOType<bool, 4> { typedef int32_t Type; }; template <> struct IOType<bool, 8> { typedef int2 Type; }; template <> struct IOType<bool, 16> { typedef int4 Type; }; template <> struct IOType<int8_t, 1> { typedef int8_t Type; }; template <> struct IOType<int8_t, 2> { typedef int16_t Type; }; template <> struct IOType<int8_t, 4> { typedef int32_t Type; }; template <> struct IOType<int8_t, 8> { typedef int2 Type; }; template <> struct IOType<int8_t, 16> { typedef int4 Type; }; template <> struct IOType<uint8_t, 1> { typedef uint8_t Type; }; template <> struct IOType<uint8_t, 2> { typedef uint16_t Type; }; template <> struct IOType<uint8_t, 4> { typedef uint32_t Type; }; template <> struct IOType<uint8_t, 8> { typedef uint2 Type; }; template <> struct IOType<uint8_t, 16> { typedef uint4 Type; }; template <> struct IOType<int16_t, 1> { typedef int16_t Type; }; template <> struct IOType<int16_t, 2> { typedef int32_t Type; }; template <> struct IOType<int16_t, 4> { typedef int2 Type; }; template <> struct IOType<int16_t, 8> { typedef int4 Type; }; template <> struct IOType<uint16_t, 1> { typedef uint16_t Type; }; template <> struct IOType<uint16_t, 2> { typedef uint32_t Type; }; template <> struct IOType<uint16_t, 4> { typedef uint2 Type; }; template <> struct IOType<uint16_t, 8> { typedef uint4 Type; }; template <> struct IOType<__half, 1> { typedef __half Type; }; template <> struct IOType<__half, 2> { typedef __half2 Type; }; template <> struct IOType<__half, 4> { typedef uint2 Type; }; template <> struct IOType<__half, 8> { typedef uint4 Type; }; template <> struct IOType<__half2, 1> { typedef __half2 Type; }; template <> struct IOType<__half2, 2> { typedef uint2 Type; }; template <> struct IOType<__half2, 4> { typedef uint4 Type; }; template <> struct IOType<int32_t, 1> { typedef int32_t Type; }; template <> struct IOType<int32_t, 2> { typedef uint2 Type; }; template <> struct IOType<int32_t, 4> { typedef uint4 Type; }; template <> struct IOType<uint32_t, 1> { typedef uint32_t Type; }; template <> struct IOType<uint32_t, 2> { typedef uint2 Type; }; template <> struct IOType<uint32_t, 4> { typedef uint4 Type; }; template <> struct IOType<float, 1> { typedef float Type; }; template <> struct IOType<float, 2> { typedef float2 Type; }; template <> struct IOType<float, 4> { typedef float4 Type; }; template <> struct IOType<int64_t, 1> { typedef int64_t Type; }; template <> struct IOType<int64_t, 2> { typedef uint4 Type; }; template <> struct IOType<uint64_t, 1> { typedef uint64_t Type; }; template <> struct IOType<uint64_t, 2> { typedef uint4 Type; }; template <> struct IOType<unsigned long long, 1> { typedef unsigned long long Type; }; template <> struct IOType<unsigned long long, 2> { typedef uint4 Type; }; template <> struct IOType<double, 1> { typedef double Type; }; template <> struct IOType<double, 2> { typedef double2 Type; }; /** * @struct TxN_t * * @brief Internal data structure that is used to define a facade for vectorized * loads/stores across the most common POD types. The goal of his file is to * provide with CUDA programmers, an easy way to have compiler issue vectorized * load or store instructions to memory (either global or shared). Vectorized * accesses to memory are important as they'll utilize its resources * efficiently, * when compared to their non-vectorized counterparts. Obviously, for whatever * reasons if one is unable to issue such vectorized operations, one can always * fallback to using POD types. * * Concept of vectorized accesses : Threads process multiple elements * to speed up processing. These are loaded in a single read thanks * to type promotion. It is then reinterpreted as a vector elements * to perform the kernel's work. * * Caution : vectorized accesses requires input addresses to be memory aligned * according not to the input type but to the promoted type used for reading. * * Example demonstrating the use of load operations, performing math on such * loaded data and finally storing it back. * @code{.cu} * TxN_t<uint8_t,8> mydata1, mydata2; * int idx = (threadIdx.x + (blockIdx.x * blockDim.x)) * mydata1.Ratio; * mydata1.load(ptr1, idx); * mydata2.load(ptr2, idx); * #pragma unroll * for(int i=0;i<mydata1.Ratio;++i) { * mydata1.val.data[i] += mydata2.val.data[i]; * } * mydata1.store(ptr1, idx); * @endcode * * By doing as above, the interesting thing is that the code effectively remains * almost the same, in case one wants to upgrade to TxN_t<uint16_t,16> type. * Only change required is to replace variable declaration appropriately. * * Obviously, it's caller's responsibility to take care of pointer alignment! * * @tparam math_ the data-type in which the compute/math needs to happen * @tparam veclen_ the number of 'math_' types to be loaded/stored per * instruction */ template <typename math_, int veclen_> struct TxN_t { /** underlying math data type */ typedef math_ math_t; /** internal storage data type */ typedef typename IOType<math_t, veclen_>::Type io_t; /** defines the number of 'math_t' types stored by this struct */ static const int Ratio = veclen_; struct alignas(io_t) { /** the vectorized data that is used for subsequent operations */ math_t data[Ratio]; } val; __device__ auto* vectorized_data() { return reinterpret_cast<io_t*>(val.data); } ///@todo: add default constructor /** * @brief Fill the contents of this structure with a constant value * @param _val the constant to be filled */ DI void fill(math_t _val) { #pragma unroll for (int i = 0; i < Ratio; ++i) { val.data[i] = _val; } } ///@todo: how to handle out-of-bounds!!? /** * @defgroup LoadsStores Global/Shared vectored loads or stores * * @brief Perform vectored loads/stores on this structure * @tparam idx_t index data type * @param ptr base pointer from where to load (or store) the data. It must * be aligned to 'sizeof(io_t)'! * @param idx the offset from the base pointer which will be loaded * (or stored) by the current thread. This must be aligned to 'Ratio'! * * @note: In case of loads, after a successful execution, the val.data will * be populated with the desired data loaded from the pointer location. In * case of stores, the data in the val.data will be stored to that location. * @{ */ template <typename idx_t = int> DI void load(const math_t* ptr, idx_t idx) { const io_t* bptr = reinterpret_cast<const io_t*>(&ptr[idx]); *vectorized_data() = __ldg(bptr); } template <typename idx_t = int> DI void load(math_t* ptr, idx_t idx) { io_t* bptr = reinterpret_cast<io_t*>(&ptr[idx]); *vectorized_data() = *bptr; } template <typename idx_t = int> DI void store(math_t* ptr, idx_t idx) { io_t* bptr = reinterpret_cast<io_t*>(&ptr[idx]); *bptr = *vectorized_data(); } /** @} */ }; /** this is just to keep the compiler happy! */ template <typename math_> struct TxN_t<math_, 0> { typedef math_ math_t; static const int Ratio = 1; struct { math_t data[1]; } val; DI void fill(math_t _val) {} template <typename idx_t = int> DI void load(const math_t* ptr, idx_t idx) { } template <typename idx_t = int> DI void load(math_t* ptr, idx_t idx) { } template <typename idx_t = int> DI void store(math_t* ptr, idx_t idx) { } }; } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/memory_pool-inl.hpp
/* * Copyright (c) 2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <cstddef> #include <memory> #include <raft/core/detail/macros.hpp> // RAFT_INLINE_CONDITIONAL #include <rmm/mr/device/managed_memory_resource.hpp> #include <rmm/mr/device/per_device_resource.hpp> #include <rmm/mr/device/pool_memory_resource.hpp> namespace raft { /** * @defgroup memory_pool Memory Pool * @{ */ /** * @brief Get a pointer to a pooled memory resource within the scope of the lifetime of the returned * unique pointer. * * This function is useful in the code where multiple repeated allocations/deallocations are * expected. * Use case example: * @code{.cpp} * void my_func(..., size_t n, rmm::mr::device_memory_resource* mr = nullptr) { * auto pool_guard = raft::get_pool_memory_resource(mr, 2 * n * sizeof(float)); * if (pool_guard){ * RAFT_LOG_INFO("Created a pool"); * } else { * RAFT_LOG_INFO("Using the current default or explicitly passed device memory resource"); * } * rmm::device_uvector<float> x(n, stream, mr); * rmm::device_uvector<float> y(n, stream, mr); * ... * } * @endcode * Here, the new memory resource would be created within the function scope if the passed `mr` is * null and the default resource is not a pool. After the call, `mr` contains a valid memory * resource in any case. * * @param[inout] mr if not null do nothing; otherwise get the current device resource and wrap it * into a `pool_memory_resource` if necessary and return the pointer to the result. * @param initial_size if a new memory pool is created, this would be its initial size (rounded up * to 256 bytes). * * @return if a new memory pool is created, it returns a unique_ptr to it; * this managed pointer controls the lifetime of the created memory resource. */ RAFT_INLINE_CONDITIONAL std::unique_ptr<rmm::mr::device_memory_resource> get_pool_memory_resource( rmm::mr::device_memory_resource*& mr, size_t initial_size) { using pool_res_t = rmm::mr::pool_memory_resource<rmm::mr::device_memory_resource>; std::unique_ptr<pool_res_t> pool_res{}; if (mr) return pool_res; mr = rmm::mr::get_current_device_resource(); if (!dynamic_cast<pool_res_t*>(mr) && !dynamic_cast<rmm::mr::pool_memory_resource<rmm::mr::cuda_memory_resource>*>(mr) && !dynamic_cast<rmm::mr::pool_memory_resource<rmm::mr::managed_memory_resource>*>(mr)) { pool_res = std::make_unique<pool_res_t>(mr, (initial_size + 255) & (~255)); mr = pool_res.get(); } return pool_res; } /** @} */ } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/raft_explicit.hpp
/* Copyright (c) 2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once /** * @brief Prevents a function template from being implicitly instantiated * * This macro defines a function body that can be used for function template * definitions of functions that should not be implicitly instantiated. * * When the template is erroneously implicitly instantiated, it provides a * useful error message that tells the user how to avoid the implicit * instantiation. * * The error message is generated using a static assert. It is generally tricky * to have a static assert fire only when you want it, as documented in * P2593: https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2022/p2593r0.html * * We use the strategy from paragraph 1.3 here. We define a struct * `not_allowed`, whose type is dependent on the template parameters of the * enclosing function instance. We use this struct type to instantiate the * `implicit_instantiation` template class, whose value is always false. We pass * this value to static_assert. This way, the static assert only fires when the * template is instantiated, since `implicit_instantiation` cannot be * instantiated without all the types in the enclosing function template. */ #define RAFT_EXPLICIT \ { \ /* Type of `not_allowed` depends on template parameters of enclosing function. */ \ struct not_allowed {}; \ static_assert( \ raft::util::raft_explicit::implicit_instantiation<not_allowed>::value, \ "ACCIDENTAL_IMPLICIT_INSTANTIATION\n\n" \ \ "If you see this error, then you have implicitly instantiated a function\n" \ "template. To keep compile times in check, libraft has the policy of\n" \ "explicitly instantiating templates. To fix the compilation error, follow\n" \ "these steps.\n\n" \ \ "If you scroll up or down a bit, you probably saw a line like the following:\n\n" \ \ "detected during instantiation of \"void raft::foo(T) [with T=float]\" at line [..]\n\n" \ \ "Simplest temporary solution:\n\n" \ \ " Add '#undef RAFT_EXPLICIT_INSTANTIATE_ONLY' at the top of your .cpp/.cu file.\n\n" \ \ "Best solution:\n\n" \ \ " 1. Add the following line to the file include/raft/foo.hpp:\n\n" \ \ " extern template void raft::foo<double>(double);\n\n" \ \ " 2. Add the following line to the file src/raft/foo.cpp:\n\n" \ \ " template void raft::foo<double>(double)\n"); \ \ /* Function may have non-void return type. */ \ /* To prevent warnings/errors about missing returns, throw an exception. */ \ throw "raft_explicit_error"; \ } namespace raft::util::raft_explicit { /** * @brief Template that is always false * * This template is from paragraph 1.3 of P2593: * https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2022/p2593r0.html * * The value of `value` is always false, but it depends on a template parameter. */ template <typename T> struct implicit_instantiation { static constexpr bool value = false; }; } // namespace raft::util::raft_explicit
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/warp_primitives.cuh
/* * Copyright (c) 2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <stdint.h> #include <raft/core/cudart_utils.hpp> #include <raft/core/operators.hpp> #include <raft/util/cuda_dev_essentials.cuh> namespace raft { /** True CUDA alignment of a type (adapted from CUB) */ template <typename T> struct cuda_alignment { struct Pad { T val; char byte; }; static constexpr int bytes = sizeof(Pad) - sizeof(T); }; template <typename LargeT, typename UnitT> struct is_multiple { static constexpr int large_align_bytes = cuda_alignment<LargeT>::bytes; static constexpr int unit_align_bytes = cuda_alignment<UnitT>::bytes; static constexpr bool value = (sizeof(LargeT) % sizeof(UnitT) == 0) && (large_align_bytes % unit_align_bytes == 0); }; template <typename LargeT, typename UnitT> inline constexpr bool is_multiple_v = is_multiple<LargeT, UnitT>::value; /** apply a warp-wide fence (useful from Volta+ archs) */ DI void warpFence() { #if __CUDA_ARCH__ >= 700 __syncwarp(); #endif } /** warp-wide any boolean aggregator */ DI bool any(bool inFlag, uint32_t mask = 0xffffffffu) { #if CUDART_VERSION >= 9000 inFlag = __any_sync(mask, inFlag); #else inFlag = __any(inFlag); #endif return inFlag; } /** warp-wide all boolean aggregator */ DI bool all(bool inFlag, uint32_t mask = 0xffffffffu) { #if CUDART_VERSION >= 9000 inFlag = __all_sync(mask, inFlag); #else inFlag = __all(inFlag); #endif return inFlag; } /** For every thread in the warp, set the corresponding bit to the thread's flag value. */ DI uint32_t ballot(bool inFlag, uint32_t mask = 0xffffffffu) { #if CUDART_VERSION >= 9000 return __ballot_sync(mask, inFlag); #else return __ballot(inFlag); #endif } template <typename T> struct is_shuffleable { static constexpr bool value = std::is_same_v<T, int> || std::is_same_v<T, unsigned int> || std::is_same_v<T, long> || std::is_same_v<T, unsigned long> || std::is_same_v<T, long long> || std::is_same_v<T, unsigned long long> || std::is_same_v<T, float> || std::is_same_v<T, double>; }; template <typename T> inline constexpr bool is_shuffleable_v = is_shuffleable<T>::value; /** * @brief Shuffle the data inside a warp * @tparam T the data type * @param val value to be shuffled * @param srcLane lane from where to shuffle * @param width lane width * @param mask mask of participating threads (Volta+) * @return the shuffled data */ template <typename T> DI std::enable_if_t<is_shuffleable_v<T>, T> shfl(T val, int srcLane, int width = WarpSize, uint32_t mask = 0xffffffffu) { #if CUDART_VERSION >= 9000 return __shfl_sync(mask, val, srcLane, width); #else return __shfl(val, srcLane, width); #endif } /// Overload of shfl for data types not supported by the CUDA intrinsics template <typename T> DI std::enable_if_t<!is_shuffleable_v<T>, T> shfl(T val, int srcLane, int width = WarpSize, uint32_t mask = 0xffffffffu) { using UnitT = std::conditional_t<is_multiple_v<T, int>, unsigned int, std::conditional_t<is_multiple_v<T, short>, unsigned short, unsigned char>>; constexpr int n_words = sizeof(T) / sizeof(UnitT); T output; UnitT* output_alias = reinterpret_cast<UnitT*>(&output); UnitT* input_alias = reinterpret_cast<UnitT*>(&val); unsigned int shuffle_word; shuffle_word = shfl((unsigned int)input_alias[0], srcLane, width, mask); output_alias[0] = shuffle_word; #pragma unroll for (int i = 1; i < n_words; ++i) { shuffle_word = shfl((unsigned int)input_alias[i], srcLane, width, mask); output_alias[i] = shuffle_word; } return output; } /** * @brief Shuffle the data inside a warp from lower lane IDs * @tparam T the data type * @param val value to be shuffled * @param delta lower lane ID delta from where to shuffle * @param width lane width * @param mask mask of participating threads (Volta+) * @return the shuffled data */ template <typename T> DI std::enable_if_t<is_shuffleable_v<T>, T> shfl_up(T val, int delta, int width = WarpSize, uint32_t mask = 0xffffffffu) { #if CUDART_VERSION >= 9000 return __shfl_up_sync(mask, val, delta, width); #else return __shfl_up(val, delta, width); #endif } /// Overload of shfl_up for data types not supported by the CUDA intrinsics template <typename T> DI std::enable_if_t<!is_shuffleable_v<T>, T> shfl_up(T val, int delta, int width = WarpSize, uint32_t mask = 0xffffffffu) { using UnitT = std::conditional_t<is_multiple_v<T, int>, unsigned int, std::conditional_t<is_multiple_v<T, short>, unsigned short, unsigned char>>; constexpr int n_words = sizeof(T) / sizeof(UnitT); T output; UnitT* output_alias = reinterpret_cast<UnitT*>(&output); UnitT* input_alias = reinterpret_cast<UnitT*>(&val); unsigned int shuffle_word; shuffle_word = shfl_up((unsigned int)input_alias[0], delta, width, mask); output_alias[0] = shuffle_word; #pragma unroll for (int i = 1; i < n_words; ++i) { shuffle_word = shfl_up((unsigned int)input_alias[i], delta, width, mask); output_alias[i] = shuffle_word; } return output; } /** * @brief Shuffle the data inside a warp * @tparam T the data type * @param val value to be shuffled * @param laneMask mask to be applied in order to perform xor shuffle * @param width lane width * @param mask mask of participating threads (Volta+) * @return the shuffled data */ template <typename T> DI std::enable_if_t<is_shuffleable_v<T>, T> shfl_xor(T val, int laneMask, int width = WarpSize, uint32_t mask = 0xffffffffu) { #if CUDART_VERSION >= 9000 return __shfl_xor_sync(mask, val, laneMask, width); #else return __shfl_xor(val, laneMask, width); #endif } /// Overload of shfl_xor for data types not supported by the CUDA intrinsics template <typename T> DI std::enable_if_t<!is_shuffleable_v<T>, T> shfl_xor(T val, int laneMask, int width = WarpSize, uint32_t mask = 0xffffffffu) { using UnitT = std::conditional_t<is_multiple_v<T, int>, unsigned int, std::conditional_t<is_multiple_v<T, short>, unsigned short, unsigned char>>; constexpr int n_words = sizeof(T) / sizeof(UnitT); T output; UnitT* output_alias = reinterpret_cast<UnitT*>(&output); UnitT* input_alias = reinterpret_cast<UnitT*>(&val); unsigned int shuffle_word; shuffle_word = shfl_xor((unsigned int)input_alias[0], laneMask, width, mask); output_alias[0] = shuffle_word; #pragma unroll for (int i = 1; i < n_words; ++i) { shuffle_word = shfl_xor((unsigned int)input_alias[i], laneMask, width, mask); output_alias[i] = shuffle_word; } return output; } } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/cudart_utils.hpp
/* * Copyright (c) 2019-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/core/error.hpp> #include <raft/util/cuda_rt_essentials.hpp> #include <raft/util/memory_pool.hpp> #include <rmm/cuda_stream_view.hpp> #include <cuda_fp16.h> #include <cuda_runtime_api.h> #include <chrono> #include <cstdio> #include <execinfo.h> #include <iomanip> #include <iostream> #include <memory> #include <mutex> #include <string> namespace raft { /** Helper method to get to know warp size in device code */ __host__ __device__ constexpr inline int warp_size() { return 32; } __host__ __device__ constexpr inline unsigned int warp_full_mask() { return 0xffffffff; } /** * @brief A kernel grid configuration construction gadget for simple one-dimensional mapping * elements to threads. */ class grid_1d_thread_t { public: int const block_size{0}; int const num_blocks{0}; /** * @param overall_num_elements The number of elements the kernel needs to handle/process * @param num_threads_per_block The grid block size, determined according to the kernel's * specific features (amount of shared memory necessary, SM functional units use pattern etc.); * this can't be determined generically/automatically (as opposed to the number of blocks) * @param max_num_blocks_1d maximum number of blocks in 1d grid * @param elements_per_thread Typically, a single kernel thread processes more than a single * element; this affects the number of threads the grid must contain */ grid_1d_thread_t(size_t overall_num_elements, size_t num_threads_per_block, size_t max_num_blocks_1d, size_t elements_per_thread = 1) : block_size(num_threads_per_block), num_blocks( std::min((overall_num_elements + (elements_per_thread * num_threads_per_block) - 1) / (elements_per_thread * num_threads_per_block), max_num_blocks_1d)) { RAFT_EXPECTS(overall_num_elements > 0, "overall_num_elements must be > 0"); RAFT_EXPECTS(num_threads_per_block / warp_size() > 0, "num_threads_per_block / warp_size() must be > 0"); RAFT_EXPECTS(elements_per_thread > 0, "elements_per_thread must be > 0"); } }; /** * @brief A kernel grid configuration construction gadget for simple one-dimensional mapping * elements to warps. */ class grid_1d_warp_t { public: int const block_size{0}; int const num_blocks{0}; /** * @param overall_num_elements The number of elements the kernel needs to handle/process * @param num_threads_per_block The grid block size, determined according to the kernel's * specific features (amount of shared memory necessary, SM functional units use pattern etc.); * this can't be determined generically/automatically (as opposed to the number of blocks) * @param max_num_blocks_1d maximum number of blocks in 1d grid */ grid_1d_warp_t(size_t overall_num_elements, size_t num_threads_per_block, size_t max_num_blocks_1d) : block_size(num_threads_per_block), num_blocks(std::min((overall_num_elements + (num_threads_per_block / warp_size()) - 1) / (num_threads_per_block / warp_size()), max_num_blocks_1d)) { RAFT_EXPECTS(overall_num_elements > 0, "overall_num_elements must be > 0"); RAFT_EXPECTS(num_threads_per_block / warp_size() > 0, "num_threads_per_block / warp_size() must be > 0"); } }; /** * @brief A kernel grid configuration construction gadget for simple one-dimensional mapping * elements to blocks. */ class grid_1d_block_t { public: int const block_size{0}; int const num_blocks{0}; /** * @param overall_num_elements The number of elements the kernel needs to handle/process * @param num_threads_per_block The grid block size, determined according to the kernel's * specific features (amount of shared memory necessary, SM functional units use pattern etc.); * this can't be determined generically/automatically (as opposed to the number of blocks) * @param max_num_blocks_1d maximum number of blocks in 1d grid */ grid_1d_block_t(size_t overall_num_elements, size_t num_threads_per_block, size_t max_num_blocks_1d) : block_size(num_threads_per_block), num_blocks(std::min(overall_num_elements, max_num_blocks_1d)) { RAFT_EXPECTS(overall_num_elements > 0, "overall_num_elements must be > 0"); RAFT_EXPECTS(num_threads_per_block / warp_size() > 0, "num_threads_per_block / warp_size() must be > 0"); } }; /** * @brief Generic copy method for all kinds of transfers * @tparam Type data type * @param dst destination pointer * @param src source pointer * @param len length of the src/dst buffers in terms of number of elements * @param stream cuda stream */ template <typename Type> void copy(Type* dst, const Type* src, size_t len, rmm::cuda_stream_view stream) { RAFT_CUDA_TRY(cudaMemcpyAsync(dst, src, len * sizeof(Type), cudaMemcpyDefault, stream)); } /** * @defgroup Copy Copy methods * These are here along with the generic 'copy' method in order to improve * code readability using explicitly specified function names * @{ */ /** performs a host to device copy */ template <typename Type> void update_device(Type* d_ptr, const Type* h_ptr, size_t len, rmm::cuda_stream_view stream) { copy(d_ptr, h_ptr, len, stream); } /** performs a device to host copy */ template <typename Type> void update_host(Type* h_ptr, const Type* d_ptr, size_t len, rmm::cuda_stream_view stream) { copy(h_ptr, d_ptr, len, stream); } template <typename Type> void copy_async(Type* d_ptr1, const Type* d_ptr2, size_t len, rmm::cuda_stream_view stream) { RAFT_CUDA_TRY( cudaMemcpyAsync(d_ptr1, d_ptr2, len * sizeof(Type), cudaMemcpyDeviceToDevice, stream)); } /** @} */ /** * @defgroup Debug Utils for debugging host/device buffers * @{ */ template <class T, class OutStream> void print_host_vector(const char* variable_name, const T* host_mem, size_t componentsCount, OutStream& out) { out << variable_name << "=["; for (size_t i = 0; i < componentsCount; ++i) { if (i != 0) out << ","; out << host_mem[i]; } out << "];" << std::endl; } template <class T, class OutStream> void print_device_vector(const char* variable_name, const T* devMem, size_t componentsCount, OutStream& out) { auto host_mem = std::make_unique<T[]>(componentsCount); RAFT_CUDA_TRY( cudaMemcpy(host_mem.get(), devMem, componentsCount * sizeof(T), cudaMemcpyDeviceToHost)); print_host_vector(variable_name, host_mem.get(), componentsCount, out); } /** * @brief Print an array given a device or a host pointer. * * @param[in] variable_name * @param[in] ptr any pointer (device/host/managed, etc) * @param[in] componentsCount array length * @param out the output stream */ template <class T, class OutStream> void print_vector(const char* variable_name, const T* ptr, size_t componentsCount, OutStream& out) { cudaPointerAttributes attr; RAFT_CUDA_TRY(cudaPointerGetAttributes(&attr, ptr)); if (attr.hostPointer != nullptr) { print_host_vector(variable_name, reinterpret_cast<T*>(attr.hostPointer), componentsCount, out); } else if (attr.type == cudaMemoryTypeUnregistered) { print_host_vector(variable_name, ptr, componentsCount, out); } else { print_device_vector(variable_name, ptr, componentsCount, out); } } /** @} */ /** * Returns the id of the device for which the pointer is located * @param p pointer to check * @return id of device for which pointer is located, otherwise -1. */ template <typename T> int get_device_for_address(const T* p) { if (!p) { return -1; } cudaPointerAttributes att; cudaError_t err = cudaPointerGetAttributes(&att, p); if (err == cudaErrorInvalidValue) { // Make sure the current thread error status has been reset err = cudaGetLastError(); return -1; } // memoryType is deprecated for CUDA 10.0+ if (att.type == cudaMemoryTypeDevice) { return att.device; } else { return -1; } } /** helper method to get max usable shared mem per block parameter */ inline int getSharedMemPerBlock() { int devId; RAFT_CUDA_TRY(cudaGetDevice(&devId)); int smemPerBlk; RAFT_CUDA_TRY(cudaDeviceGetAttribute(&smemPerBlk, cudaDevAttrMaxSharedMemoryPerBlock, devId)); return smemPerBlk; } /** helper method to get multi-processor count parameter */ inline int getMultiProcessorCount() { int devId; RAFT_CUDA_TRY(cudaGetDevice(&devId)); int mpCount; RAFT_CUDA_TRY(cudaDeviceGetAttribute(&mpCount, cudaDevAttrMultiProcessorCount, devId)); return mpCount; } /** helper method to get major minor compute capability version */ inline std::pair<int, int> getComputeCapability() { int devId; RAFT_CUDA_TRY(cudaGetDevice(&devId)); int majorVer, minorVer; RAFT_CUDA_TRY(cudaDeviceGetAttribute(&majorVer, cudaDevAttrComputeCapabilityMajor, devId)); RAFT_CUDA_TRY(cudaDeviceGetAttribute(&minorVer, cudaDevAttrComputeCapabilityMinor, devId)); return std::make_pair(majorVer, minorVer); } /** helper method to convert an array on device to a string on host */ template <typename T> std::string arr2Str(const T* arr, int size, std::string name, cudaStream_t stream, int width = 4) { std::stringstream ss; T* arr_h = (T*)malloc(size * sizeof(T)); update_host(arr_h, arr, size, stream); RAFT_CUDA_TRY(cudaStreamSynchronize(stream)); ss << name << " = [ "; for (int i = 0; i < size; i++) { typedef typename std::conditional_t<std::is_same_v<T, int8_t> || std::is_same_v<T, uint8_t>, int, T> CastT; auto val = static_cast<CastT>(arr_h[i]); ss << std::setw(width) << val; if (i < size - 1) ss << ", "; } ss << " ]" << std::endl; free(arr_h); return ss.str(); } /** this seems to be unused, but may be useful in the future */ template <typename T> void ASSERT_DEVICE_MEM(T* ptr, std::string name) { cudaPointerAttributes s_att; cudaError_t s_err = cudaPointerGetAttributes(&s_att, ptr); if (s_err != 0 || s_att.device == -1) std::cout << "Invalid device pointer encountered in " << name << ". device=" << s_att.device << ", err=" << s_err << std::endl; } inline uint32_t curTimeMillis() { auto now = std::chrono::high_resolution_clock::now(); auto duration = now.time_since_epoch(); return std::chrono::duration_cast<std::chrono::milliseconds>(duration).count(); } /** Helper function to calculate need memory for allocate to store dense matrix. * @param rows number of rows in matrix * @param columns number of columns in matrix * @return need number of items to allocate via allocate() * @sa allocate() */ inline size_t allocLengthForMatrix(size_t rows, size_t columns) { return rows * columns; } /** Helper function to check alignment of pointer. * @param ptr the pointer to check * @param alignment to be checked for * @return true if address in bytes is a multiple of alignment */ template <typename Type> bool is_aligned(Type* ptr, size_t alignment) { return reinterpret_cast<uintptr_t>(ptr) % alignment == 0; } /** calculate greatest common divisor of two numbers * @a integer * @b integer * @ return gcd of a and b */ template <typename IntType> constexpr IntType gcd(IntType a, IntType b) { while (b != 0) { IntType tmp = b; b = a % b; a = tmp; } return a; } template <typename T> constexpr T lower_bound() { if constexpr (std::numeric_limits<T>::has_infinity && std::numeric_limits<T>::is_signed) { return -std::numeric_limits<T>::infinity(); } return std::numeric_limits<T>::lowest(); } template <typename T> constexpr T upper_bound() { if constexpr (std::numeric_limits<T>::has_infinity) { return std::numeric_limits<T>::infinity(); } return std::numeric_limits<T>::max(); } namespace { // NOLINT /** * This is a hack to allow constexpr definition of `half` constants. * * Neither union-based nor reinterpret_cast-based type punning is possible within * constexpr; at the same time, all non-default constructors of `half` data type are not constexpr * as well. * * Based on the implementation details in `cuda_fp16.hpp`, we define here a new constructor for * `half` data type, that is a proper constexpr. * * When we switch to C++20, perhaps we can use `bit_cast` for the same purpose. */ struct __half_constexpr : __half { // NOLINT constexpr explicit inline __half_constexpr(uint16_t u) : __half() { __x = u; } }; } // namespace template <> constexpr inline auto lower_bound<half>() -> half { return static_cast<half>(__half_constexpr{0xfc00u}); } template <> constexpr inline auto upper_bound<half>() -> half { return static_cast<half>(__half_constexpr{0x7c00u}); } } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/cuda_rt_essentials.hpp
/* * Copyright (c) 2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once // This file provides a few essential functions that wrap the CUDA runtime API. // The scope is necessarily limited to ensure that compilation times are // minimized. Please make sure not to include large / expensive files from here. #include <cuda_runtime.h> #include <raft/core/error.hpp> #include <cstdio> namespace raft { /** * @brief Exception thrown when a CUDA error is encountered. */ struct cuda_error : public raft::exception { explicit cuda_error(char const* const message) : raft::exception(message) {} explicit cuda_error(std::string const& message) : raft::exception(message) {} }; } // namespace raft /** * @brief Error checking macro for CUDA runtime API functions. * * Invokes a CUDA runtime API function call, if the call does not return * cudaSuccess, invokes cudaGetLastError() to clear the error and throws an * exception detailing the CUDA error that occurred * */ #define RAFT_CUDA_TRY(call) \ do { \ cudaError_t const status = call; \ if (status != cudaSuccess) { \ cudaGetLastError(); \ std::string msg{}; \ SET_ERROR_MSG(msg, \ "CUDA error encountered at: ", \ "call='%s', Reason=%s:%s", \ #call, \ cudaGetErrorName(status), \ cudaGetErrorString(status)); \ throw raft::cuda_error(msg); \ } \ } while (0) /** * @brief Debug macro to check for CUDA errors * * In a non-release build, this macro will synchronize the specified stream * before error checking. In both release and non-release builds, this macro * checks for any pending CUDA errors from previous calls. If an error is * reported, an exception is thrown detailing the CUDA error that occurred. * * The intent of this macro is to provide a mechanism for synchronous and * deterministic execution for debugging asynchronous CUDA execution. It should * be used after any asynchronous CUDA call, e.g., cudaMemcpyAsync, or an * asynchronous kernel launch. */ #ifndef NDEBUG #define RAFT_CHECK_CUDA(stream) RAFT_CUDA_TRY(cudaStreamSynchronize(stream)); #else #define RAFT_CHECK_CUDA(stream) RAFT_CUDA_TRY(cudaPeekAtLastError()); #endif // /** // * @brief check for cuda runtime API errors but log error instead of raising // * exception. // */ #define RAFT_CUDA_TRY_NO_THROW(call) \ do { \ cudaError_t const status = call; \ if (cudaSuccess != status) { \ printf("CUDA call='%s' at file=%s line=%d failed with %s\n", \ #call, \ __FILE__, \ __LINE__, \ cudaGetErrorString(status)); \ } \ } while (0)
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/reduction.cuh
/* * Copyright (c) 2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <stdint.h> #include <raft/core/cudart_utils.hpp> #include <raft/core/operators.hpp> #include <raft/util/cuda_dev_essentials.cuh> #include <raft/util/warp_primitives.cuh> namespace raft { /** * @brief Logical-warp-level reduction * @tparam logicalWarpSize Logical warp size (2, 4, 8, 16 or 32) * @tparam T Value type to be reduced * @tparam ReduceLambda Reduction operation type * @param val input value * @param reduce_op Reduction operation * @return Reduction result. All lanes will have the valid result. */ template <int logicalWarpSize, typename T, typename ReduceLambda> DI T logicalWarpReduce(T val, ReduceLambda reduce_op) { #pragma unroll for (int i = logicalWarpSize / 2; i > 0; i >>= 1) { T tmp = shfl_xor(val, i); val = reduce_op(val, tmp); } return val; } /** * @brief Warp-level reduction * @tparam T Value type to be reduced * @tparam ReduceLambda Reduction operation type * @param val input value * @param reduce_op Reduction operation * @return Reduction result. All lanes will have the valid result. * @note Why not cub? Because cub doesn't seem to allow working with arbitrary * number of warps in a block. All threads in the warp must enter this * function together */ template <typename T, typename ReduceLambda> DI T warpReduce(T val, ReduceLambda reduce_op) { return logicalWarpReduce<WarpSize>(val, reduce_op); } /** * @brief Warp-level reduction * @tparam T Value type to be reduced * @param val input value * @return Reduction result. All lanes will have the valid result. * @note Why not cub? Because cub doesn't seem to allow working with arbitrary * number of warps in a block. All threads in the warp must enter this * function together */ template <typename T> DI T warpReduce(T val) { return warpReduce(val, raft::add_op{}); } /** * @brief 1-D block-level reduction * @param val input value * @param smem shared memory region needed for storing intermediate results. It * must alteast be of size: `sizeof(T) * nWarps` * @param reduce_op a binary reduction operation. * @return only the thread0 will contain valid reduced result * @note Why not cub? Because cub doesn't seem to allow working with arbitrary * number of warps in a block. All threads in the block must enter this * function together. cub also uses too many registers */ template <typename T, typename ReduceLambda = raft::add_op> DI T blockReduce(T val, char* smem, ReduceLambda reduce_op = raft::add_op{}) { auto* sTemp = reinterpret_cast<T*>(smem); int nWarps = (blockDim.x + WarpSize - 1) / WarpSize; int lid = laneId(); int wid = threadIdx.x / WarpSize; val = warpReduce(val, reduce_op); if (lid == 0) sTemp[wid] = val; __syncthreads(); val = lid < nWarps ? sTemp[lid] : T(0); return warpReduce(val, reduce_op); } /** * @brief 1-D warp-level ranked reduction which returns the value and rank. * thread 0 will have valid result and rank(idx). * @param val input value * @param idx index to be used as rank * @param reduce_op a binary reduction operation. */ template <typename T, typename ReduceLambda, typename i_t = int> DI void warpRankedReduce(T& val, i_t& idx, ReduceLambda reduce_op = raft::min_op{}) { #pragma unroll for (i_t offset = WarpSize / 2; offset > 0; offset /= 2) { T tmpVal = shfl(val, laneId() + offset); i_t tmpIdx = shfl(idx, laneId() + offset); if (reduce_op(tmpVal, val) == tmpVal) { val = tmpVal; idx = tmpIdx; } } } /** * @brief 1-D block-level ranked reduction which returns the value and rank. * thread 0 will have valid result and rank(idx). * @param val input value * @param shbuf shared memory region needed for storing intermediate results. It * must alteast be of size: `(sizeof(T) + sizeof(i_t)) * WarpSize` * @param idx index to be used as rank * @param reduce_op binary min or max operation. * @return only the thread0 will contain valid reduced result */ template <typename T, typename ReduceLambda, typename i_t = int> DI std::pair<T, i_t> blockRankedReduce(T val, T* shbuf, i_t idx = threadIdx.x, ReduceLambda reduce_op = raft::min_op{}) { T* values = shbuf; i_t* indices = (i_t*)&shbuf[WarpSize]; i_t wid = threadIdx.x / WarpSize; i_t nWarps = (blockDim.x + WarpSize - 1) / WarpSize; warpRankedReduce(val, idx, reduce_op); // Each warp performs partial reduction i_t lane = laneId(); if (lane == 0) { values[wid] = val; // Write reduced value to shared memory indices[wid] = idx; // Write reduced value to shared memory } __syncthreads(); // Wait for all partial reductions // read from shared memory only if that warp existed if (lane < nWarps) { val = values[lane]; idx = indices[lane]; } else { // get the min if it is a max op, get the max if it is a min op val = reduce_op(std::numeric_limits<T>::min(), std::numeric_limits<T>::max()) == std::numeric_limits<T>::min() ? std::numeric_limits<T>::max() : std::numeric_limits<T>::min(); idx = -1; } __syncthreads(); if (wid == 0) warpRankedReduce(val, idx, reduce_op); return std::pair<T, i_t>{val, idx}; } /** * @brief Executes a 1d binary block reduce * @param val binary value to be reduced across the thread block * @param shmem memory needed for the reduction. It should be at least of size blockDim.x/WarpSize * @return only the thread0 will contain valid reduced result */ template <int BLOCK_SIZE, typename i_t> DI i_t binaryBlockReduce(i_t val, i_t* shmem) { static_assert(BLOCK_SIZE <= 1024); assert(val == 0 || val == 1); const uint32_t mask = __ballot_sync(~0, val); const uint32_t n_items = __popc(mask); // Each first thread of the warp if (threadIdx.x % WarpSize == 0) { shmem[threadIdx.x / WarpSize] = n_items; } __syncthreads(); val = (threadIdx.x < BLOCK_SIZE / WarpSize) ? shmem[threadIdx.x] : 0; if (threadIdx.x < WarpSize) { return warpReduce(val); } // Only first warp gets the results else { return -1; } } } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/util/cuda_dev_essentials.cuh
/* * Copyright (c) 2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once // This file provides a few essential functions for use in __device__ code. The // scope is necessarily limited to ensure that compilation times are minimized. // Please make sure not to include large / expensive files from here. namespace raft { /** helper macro for device inlined functions */ #define DI inline __device__ #define HDI inline __host__ __device__ #define HD __host__ __device__ /** * @brief Provide a ceiling division operation ie. ceil(a / b) * @tparam IntType supposed to be only integers for now! */ template <typename IntType> constexpr HDI IntType ceildiv(IntType a, IntType b) { return (a + b - 1) / b; } /** * @brief Provide an alignment function ie. ceil(a / b) * b * @tparam IntType supposed to be only integers for now! */ template <typename IntType> constexpr HDI IntType alignTo(IntType a, IntType b) { return ceildiv(a, b) * b; } /** * @brief Provide an alignment function ie. (a / b) * b * @tparam IntType supposed to be only integers for now! */ template <typename IntType> constexpr HDI IntType alignDown(IntType a, IntType b) { return (a / b) * b; } /** * @brief Check if the input is a power of 2 * @tparam IntType data type (checked only for integers) */ template <typename IntType> constexpr HDI bool isPo2(IntType num) { return (num && !(num & (num - 1))); } /** * @brief Give logarithm of the number to base-2 * @tparam IntType data type (checked only for integers) */ template <typename IntType> constexpr HDI IntType log2(IntType num, IntType ret = IntType(0)) { return num <= IntType(1) ? ret : log2(num >> IntType(1), ++ret); } /** number of threads per warp */ static const int WarpSize = 32; /** get the laneId of the current thread */ DI int laneId() { int id; asm("mov.s32 %0, %%laneid;" : "=r"(id)); return id; } /** Device function to apply the input lambda across threads in the grid */ template <int ItemsPerThread, typename L> DI void forEach(int num, L lambda) { int idx = (blockDim.x * blockIdx.x) + threadIdx.x; const int numThreads = blockDim.x * gridDim.x; #pragma unroll for (int itr = 0; itr < ItemsPerThread; ++itr, idx += numThreads) { if (idx < num) lambda(idx, itr); } } /** * @brief Swap two values * @tparam T the datatype of the values * @param a first input * @param b second input */ template <typename T> HDI void swapVals(T& a, T& b) { T tmp = a; a = b; b = tmp; } } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft/util
rapidsai_public_repos/raft/cpp/include/raft/util/detail/itertools.hpp
/* * Copyright (c) 2022, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <tuple> #include <vector> namespace raft::util::itertools::detail { template <class S, typename... Args, size_t... Is> inline std::vector<S> product(std::index_sequence<Is...> index, const std::vector<Args>&... vecs) { size_t len = 1; ((len *= vecs.size()), ...); std::vector<S> out; out.reserve(len); for (size_t i = 0; i < len; i++) { std::tuple<Args...> tup; size_t mod = len, new_mod; ((new_mod = mod / vecs.size(), std::get<Is>(tup) = vecs[(i % mod) / new_mod], mod = new_mod), ...); out.push_back({std::get<Is>(tup)...}); } return out; } } // namespace raft::util::itertools::detail
0
rapidsai_public_repos/raft/cpp/include/raft/util
rapidsai_public_repos/raft/cpp/include/raft/util/detail/cub_wrappers.cuh
/* * Copyright (c) 2019-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <cub/cub.cuh> #include <rmm/device_uvector.hpp> namespace raft { /** * @brief Convenience wrapper over cub's SortPairs method * @tparam KeyT key type * @tparam ValueT value type * @param workspace workspace buffer which will get resized if not enough space * @param inKeys input keys array * @param outKeys output keys array * @param inVals input values array * @param outVals output values array * @param len array length * @param stream cuda stream */ template <typename KeyT, typename ValueT> void sortPairs(rmm::device_uvector<char>& workspace, const KeyT* inKeys, KeyT* outKeys, const ValueT* inVals, ValueT* outVals, int len, cudaStream_t stream) { size_t worksize = 0; // Fix 'worksize' may be used uninitialized in this function. cub::DeviceRadixSort::SortPairs( nullptr, worksize, inKeys, outKeys, inVals, outVals, len, 0, sizeof(KeyT) * 8, stream); workspace.resize(worksize, stream); cub::DeviceRadixSort::SortPairs( workspace.data(), worksize, inKeys, outKeys, inVals, outVals, len, 0, sizeof(KeyT) * 8, stream); } } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft/util
rapidsai_public_repos/raft/cpp/include/raft/util/detail/scatter.cuh
/* * Copyright (c) 2019-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/util/cuda_utils.cuh> #include <raft/util/vectorized.cuh> namespace raft::detail { template <typename DataT, int VecLen, typename Lambda, typename IdxT> RAFT_KERNEL scatterKernel(DataT* out, const DataT* in, const IdxT* idx, IdxT len, Lambda op) { typedef TxN_t<DataT, VecLen> DataVec; typedef TxN_t<IdxT, VecLen> IdxVec; IdxT tid = threadIdx.x + ((IdxT)blockIdx.x * blockDim.x); tid *= VecLen; if (tid >= len) return; IdxVec idxIn; idxIn.load(idx, tid); DataVec dataIn; #pragma unroll for (int i = 0; i < VecLen; ++i) { auto inPos = idxIn.val.data[i]; dataIn.val.data[i] = op(in[inPos], tid + i); } dataIn.store(out, tid); } template <typename DataT, int VecLen, typename Lambda, typename IdxT, int TPB> void scatterImpl( DataT* out, const DataT* in, const IdxT* idx, IdxT len, Lambda op, cudaStream_t stream) { const IdxT nblks = raft::ceildiv(VecLen ? len / VecLen : len, (IdxT)TPB); scatterKernel<DataT, VecLen, Lambda, IdxT><<<nblks, TPB, 0, stream>>>(out, in, idx, len, op); RAFT_CUDA_TRY(cudaGetLastError()); } } // namespace raft::detail
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/solver/linear_assignment.cuh
/* * Copyright (c) 2020-2023, NVIDIA CORPORATION. * Copyright 2020 KETAN DATE & RAKESH NAGI * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. * * CUDA Implementation of O(n^3) alternating tree Hungarian Algorithm * Authors: Ketan Date and Rakesh Nagi * * Article reference: * Date, Ketan, and Rakesh Nagi. "GPU-accelerated Hungarian algorithms * for the Linear Assignment Problem." Parallel Computing 57 (2016): 52-72. * */ #ifndef __LAP_H #define __LAP_H #pragma once #include <raft/core/resource/cuda_stream.hpp> #include <raft/core/resources.hpp> #include <rmm/device_uvector.hpp> #include <thrust/execution_policy.h> #include <thrust/fill.h> #include <raft/solver/detail/lap_functions.cuh> #include <raft/solver/linear_assignment_types.hpp> namespace raft::solver { /** * @brief CUDA Implementation of O(n^3) alternating tree Hungarian Algorithm * @note This is a port to RAFT from original authors Ketan Date and Rakesh Nagi * * @see Date, Ketan, and Rakesh Nagi. "GPU-accelerated Hungarian algorithms * for the Linear Assignment Problem." Parallel Computing 57 (2016): 52-72. * * @tparam vertex_t * @tparam weight_t */ template <typename vertex_t, typename weight_t> class LinearAssignmentProblem { private: vertex_t size_; vertex_t batchsize_; weight_t epsilon_; weight_t const* d_costs_; Vertices<vertex_t, weight_t> d_vertices_dev; VertexData<vertex_t> d_row_data_dev, d_col_data_dev; raft::resources const& handle_; rmm::device_uvector<int> row_covers_v; rmm::device_uvector<int> col_covers_v; rmm::device_uvector<weight_t> row_duals_v; rmm::device_uvector<weight_t> col_duals_v; rmm::device_uvector<weight_t> col_slacks_v; rmm::device_uvector<int> row_is_visited_v; rmm::device_uvector<int> col_is_visited_v; rmm::device_uvector<vertex_t> row_parents_v; rmm::device_uvector<vertex_t> col_parents_v; rmm::device_uvector<vertex_t> row_children_v; rmm::device_uvector<vertex_t> col_children_v; rmm::device_uvector<weight_t> obj_val_primal_v; rmm::device_uvector<weight_t> obj_val_dual_v; public: /** * @brief Constructor * @param handle raft handle for managing resources * @param size size of square matrix * @param batchsize * @param epsilon */ LinearAssignmentProblem(raft::resources const& handle, vertex_t size, vertex_t batchsize, weight_t epsilon) : handle_(handle), size_(size), batchsize_(batchsize), epsilon_(epsilon), d_costs_(nullptr), row_covers_v(0, resource::get_cuda_stream(handle_)), col_covers_v(0, resource::get_cuda_stream(handle_)), row_duals_v(0, resource::get_cuda_stream(handle_)), col_duals_v(0, resource::get_cuda_stream(handle_)), col_slacks_v(0, resource::get_cuda_stream(handle_)), row_is_visited_v(0, resource::get_cuda_stream(handle_)), col_is_visited_v(0, resource::get_cuda_stream(handle_)), row_parents_v(0, resource::get_cuda_stream(handle_)), col_parents_v(0, resource::get_cuda_stream(handle_)), row_children_v(0, resource::get_cuda_stream(handle_)), col_children_v(0, resource::get_cuda_stream(handle_)), obj_val_primal_v(0, resource::get_cuda_stream(handle_)), obj_val_dual_v(0, resource::get_cuda_stream(handle_)) { } /** * Executes Hungarian algorithm on the input cost matrix. * @param d_cost_matrix * @param d_row_assignment * @param d_col_assignment */ void solve(weight_t const* d_cost_matrix, vertex_t* d_row_assignment, vertex_t* d_col_assignment) { initializeDevice(); d_vertices_dev.row_assignments = d_row_assignment; d_vertices_dev.col_assignments = d_col_assignment; d_costs_ = d_cost_matrix; int step = 0; while (step != 100) { switch (step) { case 0: step = hungarianStep0(); break; case 1: step = hungarianStep1(); break; case 2: step = hungarianStep2(); break; case 3: step = hungarianStep3(); break; case 4: step = hungarianStep4(); break; case 5: step = hungarianStep5(); break; case 6: step = hungarianStep6(); break; } } d_costs_ = nullptr; } /** * Function for getting optimal row dual vector for subproblem spId. * @param spId * @return */ std::pair<const weight_t*, vertex_t> getRowDualVector(int spId) const { return std::make_pair(row_duals_v.data() + spId * size_, size_); } /** * Function for getting optimal col dual vector for subproblem spId. * @param spId * @return */ std::pair<const weight_t*, vertex_t> getColDualVector(int spId) { return std::make_pair(col_duals_v.data() + spId * size_, size_); } /** * Function for getting optimal primal objective value for subproblem spId. * @param spId * @return */ weight_t getPrimalObjectiveValue(int spId) { weight_t result; raft::update_host( &result, obj_val_primal_v.data() + spId, 1, resource::get_cuda_stream(handle_)); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle_)); return result; } /** * Function for getting optimal dual objective value for subproblem spId. * @param spId * @return */ weight_t getDualObjectiveValue(int spId) { weight_t result; raft::update_host(&result, obj_val_dual_v.data() + spId, 1, resource::get_cuda_stream(handle_)); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle_)); return result; } private: // Helper function for initializing global variables and arrays on a single host. void initializeDevice() { cudaStream_t stream = resource::get_cuda_stream(handle_); row_covers_v.resize(batchsize_ * size_, stream); col_covers_v.resize(batchsize_ * size_, stream); row_duals_v.resize(batchsize_ * size_, stream); col_duals_v.resize(batchsize_ * size_, stream); col_slacks_v.resize(batchsize_ * size_, stream); row_is_visited_v.resize(batchsize_ * size_, stream); col_is_visited_v.resize(batchsize_ * size_, stream); row_parents_v.resize(batchsize_ * size_, stream); col_parents_v.resize(batchsize_ * size_, stream); row_children_v.resize(batchsize_ * size_, stream); col_children_v.resize(batchsize_ * size_, stream); obj_val_primal_v.resize(batchsize_, stream); obj_val_dual_v.resize(batchsize_, stream); d_vertices_dev.row_covers = row_covers_v.data(); d_vertices_dev.col_covers = col_covers_v.data(); d_vertices_dev.row_duals = row_duals_v.data(); d_vertices_dev.col_duals = col_duals_v.data(); d_vertices_dev.col_slacks = col_slacks_v.data(); d_row_data_dev.is_visited = row_is_visited_v.data(); d_col_data_dev.is_visited = col_is_visited_v.data(); d_row_data_dev.parents = row_parents_v.data(); d_row_data_dev.children = row_children_v.data(); d_col_data_dev.parents = col_parents_v.data(); d_col_data_dev.children = col_children_v.data(); thrust::fill(thrust::device, row_covers_v.begin(), row_covers_v.end(), int{0}); thrust::fill(thrust::device, col_covers_v.begin(), col_covers_v.end(), int{0}); thrust::fill(thrust::device, row_duals_v.begin(), row_duals_v.end(), weight_t{0}); thrust::fill(thrust::device, col_duals_v.begin(), col_duals_v.end(), weight_t{0}); } // Function for calculating initial zeros by subtracting row and column minima from each element. int hungarianStep0() { detail::initialReduction(handle_, d_costs_, d_vertices_dev, batchsize_, size_); return 1; } // Function for calculating initial zeros by subtracting row and column minima from each element. int hungarianStep1() { detail::computeInitialAssignments( handle_, d_costs_, d_vertices_dev, batchsize_, size_, epsilon_); int next = 2; while (true) { if ((next = hungarianStep2()) == 6) break; if ((next = hungarianStep3()) == 5) break; hungarianStep4(); } return next; } // Function for checking optimality and constructing predicates and covers. int hungarianStep2() { int cover_count = detail::computeRowCovers( handle_, d_vertices_dev, d_row_data_dev, d_col_data_dev, batchsize_, size_); int next = (cover_count == batchsize_ * size_) ? 6 : 3; return next; } // Function for building alternating tree rooted at unassigned rows. int hungarianStep3() { int next; rmm::device_scalar<bool> flag_v(resource::get_cuda_stream(handle_)); bool h_flag = false; flag_v.set_value_async(h_flag, resource::get_cuda_stream(handle_)); detail::executeZeroCover(handle_, d_costs_, d_vertices_dev, d_row_data_dev, d_col_data_dev, flag_v.data(), batchsize_, size_, epsilon_); h_flag = flag_v.value(resource::get_cuda_stream(handle_)); next = h_flag ? 4 : 5; return next; } // Function for augmenting the solution along multiple node-disjoint alternating trees. int hungarianStep4() { detail::reversePass(handle_, d_row_data_dev, d_col_data_dev, batchsize_, size_); detail::augmentationPass( handle_, d_vertices_dev, d_row_data_dev, d_col_data_dev, batchsize_, size_); return 2; } // Function for updating dual solution to introduce new zero-cost arcs. int hungarianStep5() { detail::dualUpdate( handle_, d_vertices_dev, d_row_data_dev, d_col_data_dev, batchsize_, size_, epsilon_); return 3; } // Function for calculating primal and dual objective values at optimality. int hungarianStep6() { detail::calcObjValPrimal(handle_, obj_val_primal_v.data(), d_costs_, d_vertices_dev.row_assignments, batchsize_, size_); detail::calcObjValDual(handle_, obj_val_dual_v.data(), d_vertices_dev, batchsize_, size_); return 100; } }; } // namespace raft::solver #endif
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/solver/linear_assignment_types.hpp
/* * Copyright (c) 2020-2022, NVIDIA CORPORATION. * Copyright 2020 KETAN DATE & RAKESH NAGI * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. * * CUDA Implementation of O(n^3) alternating tree Hungarian Algorithm * Authors: Ketan Date and Rakesh Nagi * * Article reference: * Date, Ketan, and Rakesh Nagi. "GPU-accelerated Hungarian algorithms * for the Linear Assignment Problem." Parallel Computing 57 (2016): 52-72. * */ #pragma once namespace raft::solver { template <typename vertex_t, typename weight_t> struct Vertices { vertex_t* row_assignments; vertex_t* col_assignments; int* row_covers; int* col_covers; weight_t* row_duals; weight_t* col_duals; weight_t* col_slacks; }; template <typename vertex_t> struct VertexData { vertex_t* parents; vertex_t* children; int* is_visited; }; } // namespace raft::solver
0
rapidsai_public_repos/raft/cpp/include/raft/solver
rapidsai_public_repos/raft/cpp/include/raft/solver/detail/lap_functions.cuh
/* * Copyright (c) 2020-2023, NVIDIA CORPORATION. * Copyright 2020 KETAN DATE & RAKESH NAGI * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. * * CUDA Implementation of O(n^3) alternating tree Hungarian Algorithm * Authors: Ketan Date and Rakesh Nagi * * Article reference: * Date, Ketan, and Rakesh Nagi. "GPU-accelerated Hungarian algorithms * for the Linear Assignment Problem." Parallel Computing 57 (2016): 52-72. * */ #pragma once #include <raft/core/resource/cuda_stream.hpp> #include <raft/solver/linear_assignment_types.hpp> #include <raft/core/resources.hpp> #include <raft/solver/detail/lap_kernels.cuh> #include <raft/util/cudart_utils.hpp> #include <rmm/device_scalar.hpp> #include <rmm/device_uvector.hpp> #include <thrust/execution_policy.h> #include <thrust/fill.h> #include <thrust/reduce.h> #include <thrust/scan.h> #include <cstddef> namespace raft::solver::detail { const int BLOCKDIMX{64}; const int BLOCKDIMY{1}; // Function for calculating grid and block dimensions from the given input size. inline void calculateLinearDims(dim3& blocks_per_grid, dim3& threads_per_block, int& total_blocks, int size) { threads_per_block.x = BLOCKDIMX * BLOCKDIMY; int value = size / threads_per_block.x; if (size % threads_per_block.x > 0) value++; total_blocks = value; blocks_per_grid.x = value; } // Function for calculating grid and block dimensions from the given input size for square grid. inline void calculateSquareDims(dim3& blocks_per_grid, dim3& threads_per_block, int& total_blocks, int size) { threads_per_block.x = BLOCKDIMX; threads_per_block.y = BLOCKDIMY; int sq_size = (int)ceil(sqrt(size)); int valuex = (int)ceil((float)(sq_size) / BLOCKDIMX); int valuey = (int)ceil((float)(sq_size) / BLOCKDIMY); total_blocks = valuex * valuey; blocks_per_grid.x = valuex; blocks_per_grid.y = valuey; } // Function for calculating grid and block dimensions from the given input size for rectangular // grid. inline void calculateRectangularDims( dim3& blocks_per_grid, dim3& threads_per_block, int& total_blocks, int xsize, int ysize) { threads_per_block.x = BLOCKDIMX; threads_per_block.y = BLOCKDIMY; int valuex = xsize / threads_per_block.x; if (xsize % threads_per_block.x > 0) valuex++; int valuey = ysize / threads_per_block.y; if (ysize % threads_per_block.y > 0) valuey++; total_blocks = valuex * valuey; blocks_per_grid.x = valuex; blocks_per_grid.y = valuey; } template <typename vertex_t, typename weight_t> inline void initialReduction(raft::resources const& handle, weight_t const* d_costs, Vertices<vertex_t, weight_t>& d_vertices_dev, int SP, vertex_t N) { dim3 blocks_per_grid; dim3 threads_per_block; int total_blocks = 0; detail::calculateRectangularDims(blocks_per_grid, threads_per_block, total_blocks, N, SP); kernel_rowReduction<<<blocks_per_grid, threads_per_block, 0, resource::get_cuda_stream(handle)>>>( d_costs, d_vertices_dev.row_duals, SP, N, std::numeric_limits<weight_t>::max()); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle)); kernel_columnReduction<<<blocks_per_grid, threads_per_block, 0, resource::get_cuda_stream(handle)>>>( d_costs, d_vertices_dev.row_duals, d_vertices_dev.col_duals, SP, N, std::numeric_limits<weight_t>::max()); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle)); } template <typename vertex_t, typename weight_t> inline void computeInitialAssignments(raft::resources const& handle, weight_t const* d_costs, Vertices<vertex_t, weight_t>& d_vertices, int SP, vertex_t N, weight_t epsilon) { dim3 blocks_per_grid; dim3 threads_per_block; int total_blocks = 0; std::size_t size = SP * N; rmm::device_uvector<int> row_lock_v(size, resource::get_cuda_stream(handle)); rmm::device_uvector<int> col_lock_v(size, resource::get_cuda_stream(handle)); thrust::fill_n(thrust::device, d_vertices.row_assignments, size, -1); thrust::fill_n(thrust::device, d_vertices.col_assignments, size, -1); thrust::fill_n(thrust::device, row_lock_v.data(), size, 0); thrust::fill_n(thrust::device, col_lock_v.data(), size, 0); detail::calculateRectangularDims(blocks_per_grid, threads_per_block, total_blocks, N, SP); kernel_computeInitialAssignments<<<blocks_per_grid, threads_per_block, 0, resource::get_cuda_stream(handle)>>>( d_costs, d_vertices.row_duals, d_vertices.col_duals, d_vertices.row_assignments, d_vertices.col_assignments, row_lock_v.data(), col_lock_v.data(), SP, N, epsilon); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle)); } // Function for finding row cover on individual devices. template <typename vertex_t, typename weight_t> inline int computeRowCovers(raft::resources const& handle, Vertices<vertex_t, weight_t>& d_vertices, VertexData<vertex_t>& d_row_data, VertexData<vertex_t>& d_col_data, int SP, vertex_t N) { dim3 blocks_per_grid; dim3 threads_per_block; int total_blocks = 0; std::size_t size = SP * N; thrust::fill_n(thrust::device, d_vertices.row_covers, size, int{0}); thrust::fill_n(thrust::device, d_vertices.col_covers, size, int{0}); thrust::fill_n(thrust::device, d_vertices.col_slacks, size, std::numeric_limits<weight_t>::max()); thrust::fill_n(thrust::device, d_row_data.is_visited, size, DORMANT); thrust::fill_n(thrust::device, d_col_data.is_visited, size, DORMANT); thrust::fill_n(thrust::device, d_row_data.parents, size, vertex_t{-1}); thrust::fill_n(thrust::device, d_row_data.children, size, vertex_t{-1}); thrust::fill_n(thrust::device, d_col_data.parents, size, vertex_t{-1}); thrust::fill_n(thrust::device, d_col_data.children, size, vertex_t{-1}); detail::calculateRectangularDims(blocks_per_grid, threads_per_block, total_blocks, N, SP); kernel_computeRowCovers<<<blocks_per_grid, threads_per_block, 0, resource::get_cuda_stream(handle)>>>( d_vertices.row_assignments, d_vertices.row_covers, d_row_data.is_visited, SP, N); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle)); return thrust::reduce(thrust::device, d_vertices.row_covers, d_vertices.row_covers + size); } // Function for covering the zeros in uncovered rows and expanding the frontier. template <typename vertex_t, typename weight_t> inline void coverZeroAndExpand(raft::resources const& handle, weight_t const* d_costs_dev, vertex_t const* d_rows_csr_neighbors, vertex_t const* d_rows_csr_ptrs, Vertices<vertex_t, weight_t>& d_vertices_dev, VertexData<vertex_t>& d_row_data_dev, VertexData<vertex_t>& d_col_data_dev, bool* d_flag, int SP, vertex_t N, weight_t epsilon) { int total_blocks = 0; dim3 blocks_per_grid; dim3 threads_per_block; detail::calculateRectangularDims(blocks_per_grid, threads_per_block, total_blocks, N, SP); kernel_coverAndExpand<<<blocks_per_grid, threads_per_block, 0, resource::get_cuda_stream(handle)>>>(d_flag, d_rows_csr_ptrs, d_rows_csr_neighbors, d_costs_dev, d_vertices_dev, d_row_data_dev, d_col_data_dev, SP, N, epsilon); } template <typename vertex_t, typename weight_t> inline vertex_t zeroCoverIteration(raft::resources const& handle, weight_t const* d_costs_dev, Vertices<vertex_t, weight_t>& d_vertices_dev, VertexData<vertex_t>& d_row_data_dev, VertexData<vertex_t>& d_col_data_dev, bool* d_flag, int SP, vertex_t N, weight_t epsilon) { vertex_t M; rmm::device_uvector<vertex_t> csr_ptrs_v(0, resource::get_cuda_stream(handle)); rmm::device_uvector<vertex_t> csr_neighbors_v(0, resource::get_cuda_stream(handle)); { dim3 blocks_per_grid; dim3 threads_per_block; int total_blocks = 0; rmm::device_uvector<bool> predicates_v(SP * N, resource::get_cuda_stream(handle)); rmm::device_uvector<vertex_t> addresses_v(SP * N, resource::get_cuda_stream(handle)); thrust::fill_n(thrust::device, predicates_v.data(), SP * N, false); thrust::fill_n(thrust::device, addresses_v.data(), SP * N, vertex_t{0}); csr_ptrs_v.resize(SP + 1, resource::get_cuda_stream(handle)); thrust::fill_n(thrust::device, csr_ptrs_v.data(), (SP + 1), vertex_t{-1}); detail::calculateRectangularDims(blocks_per_grid, threads_per_block, total_blocks, N, SP); // construct predicate matrix for edges. kernel_rowPredicateConstructionCSR<<<blocks_per_grid, threads_per_block, 0, resource::get_cuda_stream(handle)>>>( predicates_v.data(), addresses_v.data(), d_row_data_dev.is_visited, SP, N); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle)); M = thrust::reduce(thrust::device, addresses_v.begin(), addresses_v.end()); thrust::exclusive_scan( thrust::device, addresses_v.begin(), addresses_v.end(), addresses_v.begin()); if (M > 0) { csr_neighbors_v.resize(M, resource::get_cuda_stream(handle)); kernel_rowScatterCSR<<<blocks_per_grid, threads_per_block, 0, resource::get_cuda_stream(handle)>>>(predicates_v.data(), addresses_v.data(), csr_neighbors_v.data(), csr_ptrs_v.data(), M, SP, N); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle)); } } if (M > 0) { coverZeroAndExpand(handle, d_costs_dev, csr_neighbors_v.data(), csr_ptrs_v.data(), d_vertices_dev, d_row_data_dev, d_col_data_dev, d_flag, SP, N, epsilon); } return M; } // Function for executing recursive zero cover. Returns the next step (Step 4 or Step 5) depending // on the presence of uncovered zeros. template <typename vertex_t, typename weight_t> inline void executeZeroCover(raft::resources const& handle, weight_t const* d_costs_dev, Vertices<vertex_t, weight_t>& d_vertices_dev, VertexData<vertex_t>& d_row_data_dev, VertexData<vertex_t>& d_col_data_dev, bool* d_flag, int SP, vertex_t N, weight_t epsilon) { vertex_t M = 1; while (M > 0) { M = zeroCoverIteration( handle, d_costs_dev, d_vertices_dev, d_row_data_dev, d_col_data_dev, d_flag, SP, N, epsilon); } } // Function for executing reverse pass of the maximum matching. template <typename vertex_t> inline void reversePass(raft::resources const& handle, VertexData<vertex_t>& d_row_data_dev, VertexData<vertex_t>& d_col_data_dev, int SP, int N) { int total_blocks = 0; dim3 blocks_per_grid; dim3 threads_per_block; std::size_t size = SP * N; detail::calculateLinearDims(blocks_per_grid, threads_per_block, total_blocks, size); rmm::device_uvector<bool> predicates_v(size, resource::get_cuda_stream(handle)); rmm::device_uvector<vertex_t> addresses_v(size, resource::get_cuda_stream(handle)); thrust::fill_n(thrust::device, predicates_v.data(), size, false); thrust::fill_n(thrust::device, addresses_v.data(), size, vertex_t{0}); // compact the reverse pass row vertices. kernel_augmentPredicateConstruction<<<blocks_per_grid, threads_per_block, 0, resource::get_cuda_stream(handle)>>>( predicates_v.data(), addresses_v.data(), d_col_data_dev.is_visited, size); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle)); // calculate total number of vertices. std::size_t csr_size = thrust::reduce(thrust::device, addresses_v.begin(), addresses_v.end()); // exclusive scan for calculating the scatter addresses. thrust::exclusive_scan( thrust::device, addresses_v.begin(), addresses_v.end(), addresses_v.begin()); if (csr_size > 0) { int total_blocks_1 = 0; dim3 blocks_per_grid_1; dim3 threads_per_block_1; detail::calculateLinearDims(blocks_per_grid_1, threads_per_block_1, total_blocks_1, csr_size); rmm::device_uvector<vertex_t> elements_v(csr_size, resource::get_cuda_stream(handle)); kernel_augmentScatter<<<blocks_per_grid, threads_per_block, 0, resource::get_cuda_stream(handle)>>>( elements_v.data(), predicates_v.data(), addresses_v.data(), size); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle)); kernel_reverseTraversal<<<blocks_per_grid_1, threads_per_block_1, 0, resource::get_cuda_stream(handle)>>>( elements_v.data(), d_row_data_dev, d_col_data_dev, csr_size); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle)); } } // Function for executing augmentation pass of the maximum matching. template <typename vertex_t, typename weight_t> inline void augmentationPass(raft::resources const& handle, Vertices<vertex_t, weight_t>& d_vertices_dev, VertexData<vertex_t>& d_row_data_dev, VertexData<vertex_t>& d_col_data_dev, int SP, int N) { int total_blocks = 0; dim3 blocks_per_grid; dim3 threads_per_block; detail::calculateLinearDims(blocks_per_grid, threads_per_block, total_blocks, SP * N); rmm::device_uvector<bool> predicates_v(SP * N, resource::get_cuda_stream(handle)); rmm::device_uvector<vertex_t> addresses_v(SP * N, resource::get_cuda_stream(handle)); thrust::fill_n(thrust::device, predicates_v.data(), SP * N, false); thrust::fill_n(thrust::device, addresses_v.data(), SP * N, vertex_t{0}); // compact the reverse pass row vertices. kernel_augmentPredicateConstruction<<<blocks_per_grid, threads_per_block, 0, resource::get_cuda_stream(handle)>>>( predicates_v.data(), addresses_v.data(), d_row_data_dev.is_visited, SP * N); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle)); // calculate total number of vertices. // TODO: should be vertex_t vertex_t row_ids_csr_size = thrust::reduce(thrust::device, addresses_v.begin(), addresses_v.end()); // exclusive scan for calculating the scatter addresses. thrust::exclusive_scan( thrust::device, addresses_v.begin(), addresses_v.end(), addresses_v.begin()); if (row_ids_csr_size > 0) { int total_blocks_1 = 0; dim3 blocks_per_grid_1; dim3 threads_per_block_1; detail::calculateLinearDims( blocks_per_grid_1, threads_per_block_1, total_blocks_1, row_ids_csr_size); rmm::device_uvector<vertex_t> elements_v(row_ids_csr_size, resource::get_cuda_stream(handle)); kernel_augmentScatter<<<blocks_per_grid, threads_per_block, 0, resource::get_cuda_stream(handle)>>>( elements_v.data(), predicates_v.data(), addresses_v.data(), vertex_t{SP * N}); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle)); kernel_augmentation<<<blocks_per_grid_1, threads_per_block_1, 0, resource::get_cuda_stream(handle)>>>(d_vertices_dev.row_assignments, d_vertices_dev.col_assignments, elements_v.data(), d_row_data_dev, d_col_data_dev, vertex_t{N}, row_ids_csr_size); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle)); } } template <typename vertex_t, typename weight_t> inline void dualUpdate(raft::resources const& handle, Vertices<vertex_t, weight_t>& d_vertices_dev, VertexData<vertex_t>& d_row_data_dev, VertexData<vertex_t>& d_col_data_dev, int SP, vertex_t N, weight_t epsilon) { dim3 blocks_per_grid; dim3 threads_per_block; int total_blocks; rmm::device_uvector<weight_t> sp_min_v(SP, resource::get_cuda_stream(handle)); detail::calculateLinearDims(blocks_per_grid, threads_per_block, total_blocks, SP); kernel_dualUpdate_1<<<blocks_per_grid, threads_per_block, 0, resource::get_cuda_stream(handle)>>>( sp_min_v.data(), d_vertices_dev.col_slacks, d_vertices_dev.col_covers, SP, N, std::numeric_limits<weight_t>::max()); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle)); detail::calculateRectangularDims(blocks_per_grid, threads_per_block, total_blocks, N, SP); kernel_dualUpdate_2<<<blocks_per_grid, threads_per_block, 0, resource::get_cuda_stream(handle)>>>( sp_min_v.data(), d_vertices_dev.row_duals, d_vertices_dev.col_duals, d_vertices_dev.col_slacks, d_vertices_dev.row_covers, d_vertices_dev.col_covers, d_row_data_dev.is_visited, d_col_data_dev.parents, SP, N, std::numeric_limits<weight_t>::max(), epsilon); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle)); } // Function for calculating optimal objective function value using dual variables. template <typename vertex_t, typename weight_t> inline void calcObjValDual(raft::resources const& handle, weight_t* d_obj_val, Vertices<vertex_t, weight_t>& d_vertices_dev, int SP, int N) { dim3 blocks_per_grid; dim3 threads_per_block; int total_blocks = 0; detail::calculateLinearDims(blocks_per_grid, threads_per_block, total_blocks, SP); kernel_calcObjValDual<<<blocks_per_grid, threads_per_block, 0, resource::get_cuda_stream(handle)>>>( d_obj_val, d_vertices_dev.row_duals, d_vertices_dev.col_duals, SP, N); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle)); } // Function for calculating optimal objective function value using dual variables. template <typename vertex_t, typename weight_t> inline void calcObjValPrimal(raft::resources const& handle, weight_t* d_obj_val, weight_t const* d_costs, vertex_t const* d_row_assignments, int SP, vertex_t N) { dim3 blocks_per_grid; dim3 threads_per_block; int total_blocks = 0; detail::calculateLinearDims(blocks_per_grid, threads_per_block, total_blocks, SP); kernel_calcObjValPrimal<<<blocks_per_grid, threads_per_block, 0, resource::get_cuda_stream(handle)>>>( d_obj_val, d_costs, d_row_assignments, SP, N); RAFT_CHECK_CUDA(resource::get_cuda_stream(handle)); } } // namespace raft::solver::detail
0
rapidsai_public_repos/raft/cpp/include/raft/solver
rapidsai_public_repos/raft/cpp/include/raft/solver/detail/lap_kernels.cuh
/* * Copyright (c) 2020-2023, NVIDIA CORPORATION. * Copyright 2020 KETAN DATE & RAKESH NAGI * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. * * CUDA Implementation of O(n^3) alternating tree Hungarian Algorithm * Authors: Ketan Date and Rakesh Nagi * * Article reference: * Date, Ketan, and Rakesh Nagi. "GPU-accelerated Hungarian algorithms * for the Linear Assignment Problem." Parallel Computing 57 (2016): 52-72. * */ #pragma once #include "../linear_assignment_types.hpp" #include <raft/core/resources.hpp> #include <raft/util/cudart_utils.hpp> #include <thrust/execution_policy.h> #include <thrust/for_each.h> #include <cstddef> namespace raft::solver::detail { const int DORMANT{0}; const int ACTIVE{1}; const int VISITED{2}; const int REVERSE{3}; const int AUGMENT{4}; const int MODIFIED{5}; template <typename weight_t> bool __device__ near_zero(weight_t w, weight_t epsilon) { return ((w > -epsilon) && (w < epsilon)); } template <> bool __device__ near_zero<int32_t>(int32_t w, int32_t epsilon) { return (w == 0); } template <> bool __device__ near_zero<int64_t>(int64_t w, int64_t epsilon) { return (w == 0); } // Device function for traversing the neighbors from start pointer to end pointer and updating the // covers. The function sets d_next to 4 if there are uncovered zeros, indicating the requirement of // Step 4 execution. template <typename vertex_t, typename weight_t> __device__ void cover_and_expand_row(weight_t const* d_elements, weight_t const* d_row_duals, weight_t const* d_col_duals, weight_t* d_col_slacks, int* d_row_covers, int* d_col_covers, vertex_t const* d_col_assignments, bool* d_flag, vertex_t* d_row_parents, vertex_t* d_col_parents, int* d_row_visited, int* d_col_visited, vertex_t rowid, int spid, int colid, vertex_t N, weight_t epsilon) { int ROWID = spid * N + rowid; int COLID = spid * N + colid; weight_t slack = d_elements[spid * N * N + rowid * N + colid] - d_row_duals[ROWID] - d_col_duals[COLID]; int nxt_rowid = d_col_assignments[COLID]; int NXT_ROWID = spid * N + nxt_rowid; if (rowid != nxt_rowid && d_col_covers[COLID] == 0) { if (slack < d_col_slacks[COLID]) { d_col_slacks[COLID] = slack; d_col_parents[COLID] = ROWID; } if (near_zero(d_col_slacks[COLID], epsilon)) { if (nxt_rowid != -1) { d_row_parents[NXT_ROWID] = COLID; // update parent info d_row_covers[NXT_ROWID] = 0; d_col_covers[COLID] = 1; if (d_row_visited[NXT_ROWID] != VISITED) d_row_visited[NXT_ROWID] = ACTIVE; } else { d_col_visited[COLID] = REVERSE; *d_flag = true; } } } d_row_visited[ROWID] = VISITED; } // Device function for traversing an alternating path from unassigned row to unassigned column. template <typename vertex_t> __device__ void __reverse_traversal(int* d_row_visited, vertex_t* d_row_children, vertex_t* d_col_children, vertex_t const* d_row_parents, vertex_t const* d_col_parents, int cur_colid) { int cur_rowid = -1; while (cur_colid != -1) { d_col_children[cur_colid] = cur_rowid; cur_rowid = d_col_parents[cur_colid]; d_row_children[cur_rowid] = cur_colid; cur_colid = d_row_parents[cur_rowid]; } d_row_visited[cur_rowid] = AUGMENT; } // Device function for augmenting the alternating path from unassigned column to unassigned row. template <typename vertex_t> __device__ void __augment(vertex_t* d_row_assignments, vertex_t* d_col_assignments, vertex_t const* d_row_children, vertex_t const* d_col_children, vertex_t cur_rowid, vertex_t N) { int cur_colid = -1; while (cur_rowid != -1) { cur_colid = d_row_children[cur_rowid]; d_row_assignments[cur_rowid] = cur_colid % N; d_col_assignments[cur_colid] = cur_rowid % N; cur_rowid = d_col_children[cur_colid]; } } // Kernel for reducing the rows by subtracting row minimum from each row element. // FIXME: Once cuda 10.2 is the standard should replace passing infinity // here with using cuda::std::numeric_limits<weight_t>::max() template <typename vertex_t, typename weight_t> RAFT_KERNEL kernel_rowReduction( weight_t const* d_costs, weight_t* d_row_duals, int SP, vertex_t N, weight_t infinity) { int spid = blockIdx.y * blockDim.y + threadIdx.y; int rowid = blockIdx.x * blockDim.x + threadIdx.x; weight_t min = infinity; if (spid < SP && rowid < N) { for (int colid = 0; colid < N; colid++) { weight_t slack = d_costs[spid * N * N + rowid * N + colid]; if (slack < min) { min = slack; } } d_row_duals[spid * N + rowid] = min; } } // Kernel for reducing the column by subtracting column minimum from each column element. // FIXME: Once cuda 10.2 is the standard should replace passing infinity // here with using cuda::std::numeric_limits<weight_t>::max() template <typename vertex_t, typename weight_t> RAFT_KERNEL kernel_columnReduction(weight_t const* d_costs, weight_t const* d_row_duals, weight_t* d_col_duals, int SP, vertex_t N, weight_t infinity) { int spid = blockIdx.y * blockDim.y + threadIdx.y; int colid = blockIdx.x * blockDim.x + threadIdx.x; weight_t min = infinity; if (spid < SP && colid < N) { for (int rowid = 0; rowid < N; rowid++) { weight_t cost = d_costs[spid * N * N + rowid * N + colid]; weight_t row_dual = d_row_duals[spid * N + rowid]; weight_t slack = cost - row_dual; if (slack < min) { min = slack; } } d_col_duals[spid * N + colid] = min; } } // Kernel for calculating initial assignments. template <typename vertex_t, typename weight_t> RAFT_KERNEL kernel_computeInitialAssignments(weight_t const* d_costs, weight_t const* d_row_duals, weight_t const* d_col_duals, vertex_t* d_row_assignments, vertex_t* d_col_assignments, int* d_row_lock, int* d_col_lock, int SP, vertex_t N, weight_t epsilon) { int spid = blockIdx.y * blockDim.y + threadIdx.y; int colid = blockIdx.x * blockDim.x + threadIdx.x; if (spid < SP && colid < N) { int overall_colid = spid * N + colid; weight_t col_dual = d_col_duals[overall_colid]; for (vertex_t rowid = 0; rowid < N; rowid++) { int overall_rowid = spid * N + rowid; if (d_col_lock[overall_colid] == 1) break; weight_t cost = d_costs[spid * N * N + rowid * N + colid]; weight_t row_dual = d_row_duals[overall_rowid]; weight_t slack = cost - row_dual - col_dual; if (near_zero(slack, epsilon)) { if (atomicCAS(&d_row_lock[overall_rowid], 0, 1) == 0) { d_row_assignments[overall_rowid] = colid; d_col_assignments[overall_colid] = rowid; d_col_lock[overall_colid] = 1; } } } } } // Kernel for populating the cover arrays and initializing alternating tree. template <typename vertex_t> RAFT_KERNEL kernel_computeRowCovers( vertex_t* d_row_assignments, int* d_row_covers, int* d_row_visited, int SP, vertex_t N) { int spid = blockIdx.y * blockDim.y + threadIdx.y; int rowid = blockIdx.x * blockDim.x + threadIdx.x; if (spid < SP && rowid < N) { int index = spid * N + rowid; if (d_row_assignments[index] != -1) { d_row_covers[index] = 1; } else { d_row_visited[index] = ACTIVE; } } } // Kernel for populating the predicate matrix for edges in row major format. template <typename vertex_t> RAFT_KERNEL kernel_rowPredicateConstructionCSR( bool* d_predicates, vertex_t* d_addresses, int* d_row_visited, int SP, vertex_t N) { int spid = blockIdx.y * blockDim.y + threadIdx.y; int rowid = blockIdx.x * blockDim.x + threadIdx.x; if (spid < SP && rowid < N) { int index = spid * N + rowid; if (d_row_visited[index] == ACTIVE) { d_predicates[index] = true; d_addresses[index] = 1; } else { d_predicates[index] = false; d_addresses[index] = 0; } } } // Kernel for scattering the edges based on the scatter addresses. template <typename vertex_t> RAFT_KERNEL kernel_rowScatterCSR(bool const* d_predicates, vertex_t const* d_addresses, vertex_t* d_neighbors, vertex_t* d_ptrs, vertex_t M, int SP, vertex_t N) { int spid = blockIdx.y * blockDim.y + threadIdx.y; int rowid = blockIdx.x * blockDim.x + threadIdx.x; if (spid < SP && rowid < N) { int index = spid * N + rowid; bool predicate = d_predicates[index]; vertex_t compid = d_addresses[index]; if (predicate) { d_neighbors[compid] = rowid; } if (rowid == 0) { d_ptrs[spid] = compid; d_ptrs[SP] = M; } } } // Kernel for finding the minimum zero cover. template <typename vertex_t, typename weight_t> RAFT_KERNEL kernel_coverAndExpand(bool* d_flag, vertex_t const* d_ptrs, vertex_t const* d_neighbors, weight_t const* d_elements, Vertices<vertex_t, weight_t> d_vertices, VertexData<vertex_t> d_row_data, VertexData<vertex_t> d_col_data, int SP, vertex_t N, weight_t epsilon) { int spid = blockIdx.y * blockDim.y + threadIdx.y; int colid = blockIdx.x * blockDim.x + threadIdx.x; // Load values into local memory if (spid < SP && colid < N) { thrust::for_each( thrust::seq, d_neighbors + d_ptrs[spid], d_neighbors + d_ptrs[spid + 1], [d_elements, d_vertices, d_flag, d_row_data, d_col_data, spid, colid, N, epsilon] __device__( vertex_t rowid) { cover_and_expand_row(d_elements, d_vertices.row_duals, d_vertices.col_duals, d_vertices.col_slacks, d_vertices.row_covers, d_vertices.col_covers, d_vertices.col_assignments, d_flag, d_row_data.parents, d_col_data.parents, d_row_data.is_visited, d_col_data.is_visited, rowid, spid, colid, N, epsilon); }); } } // Kernel for constructing the predicates for reverse pass or augmentation candidates. template <typename vertex_t> RAFT_KERNEL kernel_augmentPredicateConstruction(bool* d_predicates, vertex_t* d_addresses, int* d_visited, int size) { int id = blockIdx.x * blockDim.x + threadIdx.x; if (id < size) { int visited = d_visited[id]; if ((visited == REVERSE) || (visited == AUGMENT)) { d_predicates[id] = true; d_addresses[id] = 1; } else { d_predicates[id] = false; d_addresses[id] = 0; } } } // Kernel for scattering the vertices based on the scatter addresses. template <typename vertex_t> RAFT_KERNEL kernel_augmentScatter(vertex_t* d_elements, bool const* d_predicates, vertex_t const* d_addresses, std::size_t size) { int id = blockIdx.x * blockDim.x + threadIdx.x; if (id < size) { if (d_predicates[id]) { d_elements[d_addresses[id]] = id; } } } // Kernel for executing the reverse pass of the maximum matching algorithm. template <typename vertex_t> RAFT_KERNEL kernel_reverseTraversal(vertex_t* d_elements, VertexData<vertex_t> d_row_data, VertexData<vertex_t> d_col_data, int size) { int id = blockIdx.x * blockDim.x + threadIdx.x; if (id < size) { __reverse_traversal(d_row_data.is_visited, d_row_data.children, d_col_data.children, d_row_data.parents, d_col_data.parents, d_elements[id]); } } // Kernel for executing the augmentation pass of the maximum matching algorithm. template <typename vertex_t> RAFT_KERNEL kernel_augmentation(vertex_t* d_row_assignments, vertex_t* d_col_assignments, vertex_t const* d_row_elements, VertexData<vertex_t> d_row_data, VertexData<vertex_t> d_col_data, vertex_t N, vertex_t size) { int id = blockIdx.x * blockDim.x + threadIdx.x; if (id < size) { __augment(d_row_assignments, d_col_assignments, d_row_data.children, d_col_data.children, d_row_elements[id], N); } } // Kernel for updating the dual values in Step 5. // FIXME: Once cuda 10.2 is the standard should replace passing infinity // here with using cuda::std::numeric_limits<weight_t>::max() template <typename vertex_t, typename weight_t> RAFT_KERNEL kernel_dualUpdate_1(weight_t* d_sp_min, weight_t const* d_col_slacks, int const* d_col_covers, int SP, vertex_t N, weight_t infinity) { int spid = blockIdx.x * blockDim.x + threadIdx.x; if (spid < SP) { weight_t min = infinity; for (int colid = 0; colid < N; colid++) { int index = spid * N + colid; weight_t slack = d_col_slacks[index]; int col_cover = d_col_covers[index]; if (col_cover == 0) if (slack < min) min = slack; } d_sp_min[spid] = min; } } // Kernel for updating the dual values in Step 5. // FIXME: Once cuda 10.2 is the standard should replace passing infinity // here with using cuda::std::numeric_limits<weight_t>::max() template <typename vertex_t, typename weight_t> RAFT_KERNEL kernel_dualUpdate_2(weight_t const* d_sp_min, weight_t* d_row_duals, weight_t* d_col_duals, weight_t* d_col_slacks, int const* d_row_covers, int const* d_col_covers, int* d_row_visited, vertex_t* d_col_parents, int SP, vertex_t N, weight_t infinity, weight_t epsilon) { int spid = blockIdx.y * blockDim.y + threadIdx.y; int id = blockIdx.x * blockDim.x + threadIdx.x; if (spid < SP && id < N) { int index = spid * N + id; if (d_sp_min[spid] < infinity) { weight_t theta = d_sp_min[spid]; int row_cover = d_row_covers[index]; int col_cover = d_col_covers[index]; if (row_cover == 0) // Row vertex is reachable from source. d_row_duals[index] += theta; if (col_cover == 1) // Col vertex is reachable from source. d_col_duals[index] -= theta; else { // Col vertex is unreachable from source. d_col_slacks[index] -= d_sp_min[spid]; if (near_zero(d_col_slacks[index], epsilon)) { int par_rowid = d_col_parents[index]; if (par_rowid != -1) d_row_visited[par_rowid] = ACTIVE; } } } } } // Kernel for calculating optimal objective function value using dual variables. template <typename vertex_t, typename weight_t> RAFT_KERNEL kernel_calcObjValDual(weight_t* d_obj_val_dual, weight_t const* d_row_duals, weight_t const* d_col_duals, int SP, vertex_t N) { int spid = blockIdx.x * blockDim.x + threadIdx.x; if (spid < SP) { float val = 0; for (int i = 0; i < N; i++) val += (d_row_duals[spid * N + i] + d_col_duals[spid * N + i]); d_obj_val_dual[spid] = val; } } // Kernel for calculating optimal objective function value using dual variables. template <typename vertex_t, typename weight_t> RAFT_KERNEL kernel_calcObjValPrimal(weight_t* d_obj_val_primal, weight_t const* d_costs, vertex_t const* d_row_assignments, int SP, vertex_t N) { int spid = blockIdx.x * blockDim.x + threadIdx.x; if (spid < SP) { weight_t val = 0; for (int i = 0; i < N; i++) { vertex_t j = d_row_assignments[spid * N + i]; val += d_costs[spid * N * N + i * N + j]; } d_obj_val_primal[spid] = val; } } } // namespace raft::solver::detail
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/core/mdspan.hpp
/* * Copyright (c) 2022-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/core/error.hpp> #include <raft/core/mdspan_types.hpp> #include <raft/core/memory_type.hpp> #include <raft/core/detail/macros.hpp> #include <raft/core/detail/mdspan_util.cuh> #include <raft/core/host_device_accessor.hpp> #include <raft/thirdparty/mdspan/include/experimental/mdspan> namespace raft { template <typename ElementType, typename Extents, typename LayoutPolicy = layout_c_contiguous, typename AccessorPolicy = std::experimental::default_accessor<ElementType>> using mdspan = std::experimental::mdspan<ElementType, Extents, LayoutPolicy, AccessorPolicy>; namespace detail { // keeping ByteAlignment as optional to allow testing template <class ValueType, size_t ByteAlignment = 128> struct padding { static_assert(std::is_same<std::remove_cv_t<ValueType>, ValueType>::value, "std::experimental::padding ValueType has to be provided without " "const or volatile specifiers."); static_assert(ByteAlignment % sizeof(ValueType) == 0 || sizeof(ValueType) % ByteAlignment == 0, "std::experimental::padding sizeof(ValueType) has to be multiple or " "divider of ByteAlignment."); static constexpr size_t value = std::max(ByteAlignment / sizeof(ValueType), 1ul); }; // alignment fixed to 128 bytes struct alignment { static constexpr size_t value = 128; }; } // namespace detail template <typename ElementType> using layout_right_padded = std::experimental::layout_right_padded< detail::padding<std::remove_cv_t<std::remove_reference_t<ElementType>>>::value>; template <typename ElementType> using layout_left_padded = std::experimental::layout_left_padded< detail::padding<std::remove_cv_t<std::remove_reference_t<ElementType>>>::value>; template <typename ElementType, typename LayoutPolicy> using enable_if_layout_padded = std::enable_if_t<std::is_same<LayoutPolicy, layout_left_padded<ElementType>>::value || std::is_same<LayoutPolicy, layout_right_padded<ElementType>>::value>; /** * Ensure all types listed in the parameter pack `Extents` are integral types. * Usage: * put it as the last nameless template parameter of a function: * `typename = ensure_integral_extents<Extents...>` */ template <typename... Extents> using ensure_integral_extents = std::enable_if_t<std::conjunction_v<std::is_integral<Extents>...>>; /** * @\brief Template checks and helpers to determine if type T is an std::mdspan * or a derived type */ template <typename ElementType, typename Extents, typename LayoutPolicy, typename AccessorPolicy> void __takes_an_mdspan_ptr(mdspan<ElementType, Extents, LayoutPolicy, AccessorPolicy>*); template <typename T, typename = void> struct is_mdspan : std::false_type {}; template <typename T> struct is_mdspan<T, std::void_t<decltype(__takes_an_mdspan_ptr(std::declval<T*>()))>> : std::true_type {}; template <typename T, typename = void> struct is_input_mdspan : std::false_type {}; template <typename T> struct is_input_mdspan<T, std::void_t<decltype(__takes_an_mdspan_ptr(std::declval<T*>()))>> : std::bool_constant<std::is_const_v<typename T::element_type>> {}; template <typename T, typename = void> struct is_output_mdspan : std::false_type {}; template <typename T> struct is_output_mdspan<T, std::void_t<decltype(__takes_an_mdspan_ptr(std::declval<T*>()))>> : std::bool_constant<not std::is_const_v<typename T::element_type>> {}; template <typename T> using is_mdspan_t = is_mdspan<std::remove_const_t<T>>; template <typename T> using is_input_mdspan_t = is_input_mdspan<T>; template <typename T> using is_output_mdspan_t = is_output_mdspan<T>; /** * @\brief Boolean to determine if variadic template types Tn are either * raft::host_mdspan/raft::device_mdspan or their derived types */ template <typename... Tn> inline constexpr bool is_mdspan_v = std::conjunction_v<is_mdspan_t<Tn>...>; template <typename... Tn> using enable_if_mdspan = std::enable_if_t<is_mdspan_v<Tn...>>; template <typename... Tn> inline constexpr bool is_input_mdspan_v = std::conjunction_v<is_input_mdspan_t<Tn>...>; template <typename... Tn> using enable_if_input_mdspan = std::enable_if_t<is_input_mdspan_v<Tn...>>; template <typename... Tn> inline constexpr bool is_output_mdspan_v = std::conjunction_v<is_output_mdspan_t<Tn>...>; template <typename... Tn> using enable_if_output_mdspan = std::enable_if_t<is_output_mdspan_v<Tn...>>; // uint division optimization inspired by the CIndexer in cupy. Division operation is // slow on both CPU and GPU, especially 64 bit integer. So here we first try to avoid 64 // bit when the index is smaller, then try to avoid division when it's exp of 2. template <typename I, typename IndexType, size_t... Extents> RAFT_INLINE_FUNCTION auto unravel_index_impl( I idx, std::experimental::extents<IndexType, Extents...> shape) { constexpr auto kRank = static_cast<int32_t>(shape.rank()); std::size_t index[shape.rank()]{0}; // NOLINT static_assert(std::is_signed<decltype(kRank)>::value, "Don't change the type without changing the for loop."); for (int32_t dim = kRank; --dim > 0;) { auto s = static_cast<std::remove_const_t<std::remove_reference_t<I>>>(shape.extent(dim)); if (s & (s - 1)) { auto t = idx / s; index[dim] = idx - t * s; idx = t; } else { // exp of 2 index[dim] = idx & (s - 1); idx >>= detail::popc(s - 1); } } index[0] = idx; return detail::arr_to_tup(index); } /** * @brief Create a raft::mdspan * @tparam ElementType the data type of the matrix elements * @tparam IndexType the index type of the extents * @tparam LayoutPolicy policy for strides and layout ordering * @tparam is_host_accessible whether the data is accessible on host * @tparam is_device_accessible whether the data is accessible on device * @param ptr Pointer to the data * @param exts dimensionality of the array (series of integers) * @return raft::mdspan */ template <typename ElementType, typename IndexType = std::uint32_t, typename LayoutPolicy = layout_c_contiguous, bool is_host_accessible = false, bool is_device_accessible = true, size_t... Extents> constexpr auto make_mdspan(ElementType* ptr, extents<IndexType, Extents...> exts) { using accessor_type = host_device_accessor< std::experimental::default_accessor<ElementType>, detail::memory_type_from_access<is_host_accessible, is_device_accessible>()>; /*using accessor_type = host_device_accessor<std::experimental::default_accessor<ElementType>, mem_type>; */ return mdspan<ElementType, decltype(exts), LayoutPolicy, accessor_type>{ptr, exts}; } /** * @brief Create a layout_stride mapping from extents and strides * @param[in] extents the dimensionality of the layout * @param[in] strides the strides between elements in the layout * @return raft::layout_stride::mapping<Extents> */ template <typename Extents, typename Strides> auto make_strided_layout(Extents extents, Strides strides) { return layout_stride::mapping<Extents>{extents, strides}; } /** * @brief Create raft::extents to specify dimensionality * * @tparam IndexType The type of each dimension of the extents * @tparam Extents Dimensions (a series of integers) * @param exts The desired dimensions * @return raft::extents */ template <typename IndexType, typename... Extents, typename = ensure_integral_extents<Extents...>> constexpr auto make_extents(Extents... exts) { return extents<IndexType, ((void)exts, dynamic_extent)...>{exts...}; } /** * @defgroup mdspan_reshape Row- or Col-norm computation * @{ */ /** * @brief Flatten raft::mdspan into a 1-dim array view * * @tparam mdspan_type Expected type raft::host_mdspan or raft::device_mdspan * @param mds raft::host_mdspan or raft::device_mdspan object * @return raft::host_mdspan or raft::device_mdspan with vector_extent * depending on AccessoryPolicy */ template <typename mdspan_type, typename = enable_if_mdspan<mdspan_type>> auto flatten(mdspan_type mds) { RAFT_EXPECTS(mds.is_exhaustive(), "Input must be contiguous."); vector_extent<typename mdspan_type::size_type> ext{mds.size()}; return std::experimental::mdspan<typename mdspan_type::element_type, decltype(ext), typename mdspan_type::layout_type, typename mdspan_type::accessor_type>(mds.data_handle(), ext); } /** * @brief Reshape raft::host_mdspan or raft::device_mdspan * * @tparam mdspan_type Expected type raft::host_mdspan or raft::device_mdspan * @tparam IndexType the index type of the extents * @tparam Extents raft::extents for dimensions * @param mds raft::host_mdspan or raft::device_mdspan object * @param new_shape Desired new shape of the input * @return raft::host_mdspan or raft::device_mdspan, depending on AccessorPolicy */ template <typename mdspan_type, typename IndexType = std::uint32_t, size_t... Extents, typename = enable_if_mdspan<mdspan_type>> auto reshape(mdspan_type mds, extents<IndexType, Extents...> new_shape) { RAFT_EXPECTS(mds.is_exhaustive(), "Input must be contiguous."); size_t new_size = 1; for (size_t i = 0; i < new_shape.rank(); ++i) { new_size *= new_shape.extent(i); } RAFT_EXPECTS(new_size == mds.size(), "Cannot reshape array with size mismatch"); return std::experimental::mdspan<typename mdspan_type::element_type, decltype(new_shape), typename mdspan_type::layout_type, typename mdspan_type::accessor_type>(mds.data_handle(), new_shape); } /* @} */ /** * @defgroup mdspan_unravel Unravel mdspan * @{ */ /** * \brief Turns linear index into coordinate. Similar to numpy unravel_index. * * \code * auto m = make_host_matrix<float>(7, 6); * auto m_v = m.view(); * auto coord = unravel_index(2, m.extents(), typename decltype(m)::layout_type{}); * std::apply(m_v, coord) = 2; * \endcode * * \param idx The linear index. * \param shape The shape of the array to use. * \param layout Must be `layout_c_contiguous` (row-major) in current implementation. * * \return A std::tuple that represents the coordinate. */ template <typename Idx, typename IndexType, typename LayoutPolicy, size_t... Exts> RAFT_INLINE_FUNCTION auto unravel_index(Idx idx, extents<IndexType, Exts...> shape, LayoutPolicy const& layout) { static_assert(std::is_same_v<std::remove_cv_t<std::remove_reference_t<decltype(layout)>>, layout_c_contiguous>, "Only C layout is supported."); static_assert(std::is_integral_v<Idx>, "Index must be integral."); auto constexpr kIs64 = sizeof(std::remove_cv_t<std::remove_reference_t<Idx>>) == sizeof(uint64_t); if (kIs64 && static_cast<uint64_t>(idx) > std::numeric_limits<uint32_t>::max()) { return unravel_index_impl<uint64_t>(static_cast<uint64_t>(idx), shape); } else { return unravel_index_impl<uint32_t>(static_cast<uint32_t>(idx), shape); } } /** @} */ /** * @defgroup mdspan_contiguous Whether the strides imply a contiguous layout. * @{ */ /** * @brief Whether the strides imply a c-contiguous layout. */ template <typename Extents, typename Strides> [[nodiscard]] auto is_c_contiguous(Extents const& extents, Strides const& strides) -> bool { typename Extents::index_type stride = 1; for (auto r = extents.rank(); r > 0; r--) { if (stride != strides[r - 1]) { return false; } stride *= extents.extent(r - 1); } return true; } /** * @brief Whether the strides imply a f-contiguous layout. */ template <typename Extents, typename Strides> [[nodiscard]] auto is_f_contiguous(Extents const& extents, Strides const& strides) -> bool { typename Extents::index_type stride = 1; for (typename Extents::rank_type r = 0; r < extents.rank(); r++) { if (stride != strides[r]) { return false; } stride *= extents.extent(r); } return true; } /** @} */ /** * @brief Const accessor specialization for default_accessor * * @tparam ElementType * @param a * @return std::experimental::default_accessor<std::add_const_t<ElementType>> */ template <class ElementType> std::experimental::default_accessor<std::add_const_t<ElementType>> accessor_of_const( std::experimental::default_accessor<ElementType> a) { return {a}; } /** * @brief Const accessor specialization for host_device_accessor * * @tparam ElementType the data type of the mdspan elements * @tparam MemType the type of memory where the elements are stored. * @param a host_device_accessor * @return host_device_accessor<std::experimental::default_accessor<std::add_const_t<ElementType>>, * MemType> */ template <class ElementType, memory_type MemType> host_device_accessor<std::experimental::default_accessor<std::add_const_t<ElementType>>, MemType> accessor_of_const(host_device_accessor<std::experimental::default_accessor<ElementType>, MemType> a) { return {a}; } /** * @defgroup mdspan_make_const Convert an mdspan to a const type * @{ */ /** * @brief Create a copy of the given mdspan with const element type * * @tparam ElementType the const-qualified data type of the mdspan elements * @tparam Extents raft::extents for dimensions * @tparam Layout policy for strides and layout ordering * @tparam Accessor Accessor policy for the input and output * @param mds raft::mdspan object * @return raft::mdspan */ template <class ElementType, class Extents, class Layout, class Accessor> auto make_const_mdspan(mdspan<ElementType, Extents, Layout, Accessor> mds) { auto acc_c = accessor_of_const(mds.accessor()); return mdspan<std::add_const_t<ElementType>, Extents, Layout, decltype(acc_c)>{ mds.data_handle(), mds.mapping(), acc_c}; } /** @} */ } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/core/operators.cuh
/* * Copyright (c) 2022, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/core/detail/macros.hpp> #include <raft/util/device_atomics.cuh> namespace raft { /** * @defgroup DeviceFunctors Commonly used device-only functors. * @{ */ struct atomic_add_op { template <typename Type> _RAFT_DEVICE _RAFT_FORCEINLINE Type operator()(Type* address, const Type& val) { return atomicAdd(address, val); } }; struct atomic_max_op { template <typename Type> _RAFT_DEVICE _RAFT_FORCEINLINE Type operator()(Type* address, const Type& val) { return atomicMax(address, val); } }; struct atomic_min_op { template <typename Type> _RAFT_DEVICE _RAFT_FORCEINLINE Type operator()(Type* address, const Type& val) { return atomicMin(address, val); } }; /** @} */ } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/core/coo_matrix.hpp
/* * Copyright (c) 2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/core/logger.hpp> #include <raft/core/resources.hpp> #include <raft/core/span.hpp> #include <raft/core/sparse_types.hpp> namespace raft { /** * \defgroup coo_matrix COO Matrix * @{ */ template <typename RowType, typename ColType, typename NZType, int is_device> class coordinate_structure_t : public sparse_structure<RowType, ColType, NZType, is_device> { public: coordinate_structure_t(RowType n_rows, ColType n_cols, NZType nnz) : sparse_structure<RowType, ColType, NZType, is_device>(n_rows, n_cols, nnz){}; /** * Return span containing underlying rows array * @return span containing underlying rows array */ virtual span<RowType, is_device> get_rows() = 0; /** * Return span containing underlying cols array * @return span containing underlying cols array */ virtual span<ColType, is_device> get_cols() = 0; }; /** * A non-owning view into a coordinate structure * * The structure representation does not have a value/weight * component so that its const-ness can be varied from it. * * @tparam RowType * @tparam ColType */ template <typename RowType, typename ColType, typename NZType, bool is_device> class coordinate_structure_view : public coordinate_structure_t<RowType, ColType, NZType, is_device> { public: static constexpr SparsityType sparsity_type = PRESERVING; using view_type = coordinate_structure_view<RowType, ColType, NZType, is_device>; using row_type = typename sparse_structure<RowType, ColType, NZType, is_device>::row_type; using col_type = typename sparse_structure<RowType, ColType, NZType, is_device>::col_type; using nnz_type = typename sparse_structure<RowType, ColType, NZType, is_device>::nnz_type; coordinate_structure_view(span<row_type, is_device> rows, span<col_type, is_device> cols, row_type n_rows, col_type n_cols) : coordinate_structure_t<RowType, ColType, NZType, is_device>(n_rows, n_cols, rows.size()), rows_{rows}, cols_{cols} { } /** * Return span containing underlying rows array * @return span containing underlying rows array */ span<row_type, is_device> get_rows() override { return rows_; } /** * Return span containing underlying cols array * @return span containing underlying cols array */ span<col_type, is_device> get_cols() override { return cols_; } protected: raft::span<row_type, is_device> rows_; raft::span<col_type, is_device> cols_; }; /** * Represents a sparse coordinate structure (or edge list) * which can be used to model a COO matrix. * * The structure representation does not have a value/weight * component so that its const-ness can be varied from it. * * @tparam RowType * @tparam ColType * @tparam ContainerPolicy */ template <typename RowType, typename ColType, typename NZType, bool is_device, template <typename T> typename ContainerPolicy> class coordinate_structure : public coordinate_structure_t<RowType, ColType, NZType, is_device> { public: static constexpr SparsityType sparsity_type = OWNING; using sparse_structure_type = coordinate_structure_t<RowType, ColType, NZType, is_device>; using row_type = typename sparse_structure_type::row_type; using col_type = typename sparse_structure_type::col_type; using nnz_type = typename sparse_structure_type::nnz_type; using view_type = coordinate_structure_view<row_type, col_type, nnz_type, is_device>; using row_container_policy_type = ContainerPolicy<RowType>; using col_container_policy_type = ContainerPolicy<ColType>; using row_container_type = typename row_container_policy_type::container_type; using col_container_type = typename col_container_policy_type::container_type; coordinate_structure( raft::resources const& handle, row_type n_rows, col_type n_cols, nnz_type nnz = 0) noexcept(std::is_nothrow_default_constructible_v<row_container_type>) : coordinate_structure_t<RowType, ColType, NZType, is_device>(n_rows, n_cols, nnz), cp_rows_{}, cp_cols_{}, c_rows_{cp_rows_.create(handle, 0)}, c_cols_{cp_cols_.create(handle, 0)} {}; coordinate_structure(coordinate_structure const&) noexcept( std::is_nothrow_copy_constructible_v<row_container_type>) = default; coordinate_structure(coordinate_structure&&) noexcept( std::is_nothrow_move_constructible<row_container_type>::value) = default; constexpr auto operator=(coordinate_structure const&) noexcept( std::is_nothrow_copy_assignable<row_container_type>::value) -> coordinate_structure& = default; constexpr auto operator=(coordinate_structure&&) noexcept( std::is_nothrow_move_assignable<row_container_type>::value) -> coordinate_structure& = default; ~coordinate_structure() noexcept(std::is_nothrow_destructible<row_container_type>::value) = default; /** * Return a view of the coordinate structure. Structural views are sparsity-preserving * so while the structural elements can be updated in a non-const view, the sparsity * itself (number of nonzeros) cannot be changed. * @return coordinate structure view */ view_type view() { if (this->get_nnz() == 0) { RAFT_LOG_WARN( "Cannot create coordinate_structure.view() because it has not been initialized " "(sparsity is 0)"); } auto row_span = raft::span<row_type, is_device>(c_rows_.data(), this->get_nnz()); auto col_span = raft::span<col_type, is_device>(c_cols_.data(), this->get_nnz()); return view_type(row_span, col_span, this->get_n_rows(), this->get_n_cols()); } /** * Return span containing underlying rows array * @return span containing underlying rows array */ span<row_type, is_device> get_rows() override { return raft::span<row_type, is_device>(c_rows_.data(), this->get_n_rows()); } /** * Return span containing underlying cols array * @return span containing underlying cols array */ span<col_type, is_device> get_cols() override { return raft::span<col_type, is_device>(c_cols_.data(), this->get_n_cols()); } /** * Change the sparsity of the current compressed structure. This will * resize the underlying data arrays. * @param nnz new sparsity */ void initialize_sparsity(nnz_type nnz) { sparse_structure_type::initialize_sparsity(nnz); c_rows_.resize(nnz); c_cols_.resize(nnz); } protected: row_container_policy_type cp_rows_; col_container_policy_type cp_cols_; row_container_type c_rows_; col_container_type c_cols_; }; template <typename ElementType, typename RowType, typename ColType, typename NZType, bool is_device> class coo_matrix_view : public sparse_matrix_view<ElementType, coordinate_structure_view<RowType, ColType, NZType, is_device>, is_device> { public: using element_type = ElementType; using row_type = RowType; using col_type = ColType; using nnz_type = NZType; coo_matrix_view(raft::span<ElementType, is_device> element_span, coordinate_structure_view<RowType, ColType, NZType, is_device> structure_view) : sparse_matrix_view<ElementType, coordinate_structure_view<RowType, ColType, NZType, is_device>, is_device>(element_span, structure_view) { } }; template <typename ElementType, typename RowType, typename ColType, typename NZType, bool is_device, template <typename T> typename ContainerPolicy, SparsityType sparsity_type = SparsityType::OWNING, typename structure_type = std::conditional_t< sparsity_type == SparsityType::OWNING, coordinate_structure<RowType, ColType, NZType, is_device, ContainerPolicy>, coordinate_structure_view<RowType, ColType, NZType, is_device>>> class coo_matrix : public sparse_matrix<ElementType, structure_type, coo_matrix_view<ElementType, RowType, ColType, NZType, is_device>, is_device, ContainerPolicy> { public: using element_type = ElementType; using row_type = RowType; using col_type = ColType; using nnz_type = NZType; using structure_view_type = typename structure_type::view_type; using container_type = typename ContainerPolicy<ElementType>::container_type; using sparse_matrix_type = sparse_matrix<ElementType, structure_type, coo_matrix_view<ElementType, RowType, ColType, NZType, is_device>, is_device, ContainerPolicy>; static constexpr auto get_sparsity_type() { return sparsity_type; } template <SparsityType sparsity_type_ = get_sparsity_type(), typename = typename std::enable_if_t<sparsity_type_ == SparsityType::OWNING>> coo_matrix(raft::resources const& handle, RowType n_rows, ColType n_cols, NZType nnz = 0) noexcept(std::is_nothrow_default_constructible_v<container_type>) : sparse_matrix_type(handle, n_rows, n_cols, nnz){}; // Constructor that owns the data but not the structure template <SparsityType sparsity_type_ = get_sparsity_type(), typename = typename std::enable_if_t<sparsity_type_ == SparsityType::PRESERVING>> coo_matrix(raft::resources const& handle, structure_type structure) noexcept( std::is_nothrow_default_constructible_v<container_type>) : sparse_matrix_type(handle, structure){}; /** * Initialize the sparsity on this instance if it was not known upon construction * Please note this will resize the underlying memory buffers * @param nnz new sparsity to initialize. */ template <SparsityType sparsity_type_ = get_sparsity_type(), typename = typename std::enable_if_t<sparsity_type_ == SparsityType::OWNING>> void initialize_sparsity(NZType nnz) { sparse_matrix_type::initialize_sparsity(nnz); this->structure_.initialize_sparsity(nnz); } /** * Return a view of the structure underlying this matrix * @return */ structure_view_type structure_view() { if constexpr (get_sparsity_type() == SparsityType::OWNING) { return this->structure_.view(); } else { return this->structure_; } } }; /** @} */ } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/core/csr_matrix.hpp
/* * Copyright (c) 2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/core/logger.hpp> #include <raft/core/resources.hpp> #include <raft/core/span.hpp> #include <raft/core/sparse_types.hpp> namespace raft { /** * \defgroup csr_matrix CSR Matrix * @{ */ template <typename IndptrType, typename IndicesType, typename NZType, int is_device> class compressed_structure_t : public sparse_structure<IndptrType, IndicesType, NZType, is_device> { public: /** * Constructor when sparsity is already known * @param n_rows total number of rows in matrix * @param n_cols total number of columns in matrix * @param nnz sparsity of matrix */ compressed_structure_t(IndptrType n_rows, IndicesType n_cols, NZType nnz) : sparse_structure<IndptrType, IndicesType, NZType, is_device>(n_rows, n_cols, nnz){}; /** * Return span containing underlying indptr array * @return span containing underlying indptr array */ virtual span<IndptrType, is_device> get_indptr() = 0; /** * Return span containing underlying indices array * @return span containing underlying indices array */ virtual span<IndicesType, is_device> get_indices() = 0; }; /** * A non-owning view into a compressed sparse structure * * The structure representation does not have a value/weight * component so that its const-ness can be varied from it. * * @tparam IndptrType * @tparam IndicesType */ template <typename IndptrType, typename IndicesType, typename NZType, bool is_device> class compressed_structure_view : public compressed_structure_t<IndptrType, IndicesType, NZType, is_device> { public: using sparse_structure_type = compressed_structure_t<IndptrType, IndicesType, NZType, is_device>; using view_type = compressed_structure_view<IndptrType, IndicesType, NZType, is_device>; using indptr_type = typename sparse_structure_type::row_type; using indices_type = typename sparse_structure_type::col_type; using nnz_type = typename sparse_structure_type::nnz_type; compressed_structure_view(span<indptr_type, is_device> indptr, span<indices_type, is_device> indices, indices_type n_cols) : sparse_structure_type(indptr.size() - 1, n_cols, indices.size()), indptr_(indptr), indices_(indices) { } /** * Return span containing underlying indptr array * @return span containing underlying indptr array */ span<indptr_type, is_device> get_indptr() override { return indptr_; } /** * Return span containing underlying indices array * @return span containing underlying indices array */ span<indices_type, is_device> get_indices() override { return indices_; } protected: raft::span<indptr_type, is_device> indptr_; raft::span<indices_type, is_device> indices_; }; /** * Represents a sparse compressed structure (or adjacency list) * which can be used to model both a CSR and CSC matrix. * * The structure representation does not have a value/weight * component so that its const-ness can be varied from it. * * @tparam IndptrType * @tparam IndicesType * @tparam ContainerPolicy */ template <typename IndptrType, typename IndicesType, typename NZType, bool is_device, template <typename T> typename ContainerPolicy> class compressed_structure : public compressed_structure_t<IndptrType, IndicesType, NZType, is_device> { public: using sparse_structure_type = compressed_structure_t<IndptrType, IndicesType, NZType, is_device>; using indptr_type = typename sparse_structure_type::row_type; using indices_type = typename sparse_structure_type::col_type; using nnz_type = typename sparse_structure_type::nnz_type; using view_type = compressed_structure_view<IndptrType, IndicesType, NZType, is_device>; using indptr_container_policy_type = ContainerPolicy<IndptrType>; using indices_container_policy_type = ContainerPolicy<IndicesType>; using indptr_container_type = typename indptr_container_policy_type::container_type; using indices_container_type = typename indices_container_policy_type::container_type; constexpr compressed_structure( raft::resources const& handle, IndptrType n_rows, IndicesType n_cols, NZType nnz = 0) noexcept(std::is_nothrow_default_constructible_v<indptr_container_type>) : sparse_structure_type{n_rows, n_cols, nnz}, cp_indptr_{}, cp_indices_{}, c_indptr_{cp_indptr_.create(handle, n_rows + 1)}, c_indices_{cp_indices_.create(handle, nnz)} {}; compressed_structure(compressed_structure const&) noexcept( std::is_nothrow_copy_constructible_v<indptr_container_type>) = default; compressed_structure(compressed_structure&&) noexcept( std::is_nothrow_move_constructible<indptr_container_type>::value) = default; constexpr auto operator=(compressed_structure const&) noexcept( std::is_nothrow_copy_assignable<indptr_container_type>::value) -> compressed_structure& = default; constexpr auto operator=(compressed_structure&&) noexcept( std::is_nothrow_move_assignable<indptr_container_type>::value) -> compressed_structure& = default; /** * Return span containing underlying indptr array * @return span containing underlying indptr array */ span<IndptrType, is_device> get_indptr() override { return raft::span<IndptrType, is_device>(c_indptr_.data(), this->get_n_rows() + 1); } /** * Return span containing underlying indices array * @return span containing underlying indices array */ span<IndicesType, is_device> get_indices() override { if (this->get_nnz() == 0) { RAFT_LOG_WARN("Indices requested for structure that has uninitialized sparsity."); } return raft::span<IndicesType, is_device>(c_indices_.data(), this->get_nnz()); } ~compressed_structure() noexcept(std::is_nothrow_destructible<indptr_container_type>::value) = default; /** * Return a view of the compressed structure. Structural views are sparsity-preserving * so while the structural elements can be updated in a non-const view, the sparsity * itself (number of nonzeros) cannot be changed. * @return compressed structure view */ view_type view() { if (this->get_nnz() == 0) { RAFT_LOG_WARN( "Cannot create compressed_structure.view() because it has not been initialized (sparsity " "is 0)"); } auto indptr_span = raft::span<IndptrType, is_device>(c_indptr_.data(), this->get_n_rows() + 1); auto indices_span = raft::span<IndicesType, is_device>(c_indices_.data(), this->get_nnz()); return view_type(indptr_span, indices_span, this->get_n_cols()); } /** * Change the sparsity of the current compressed structure. This will * resize the underlying data arrays. * @param nnz new sparsity */ void initialize_sparsity(NZType nnz) override { sparse_structure_type::initialize_sparsity(nnz); c_indptr_.resize(this->get_n_rows() + 1); c_indices_.resize(nnz); } protected: indptr_container_policy_type cp_indptr_; indices_container_policy_type cp_indices_; indptr_container_type c_indptr_; indices_container_type c_indices_; }; template <typename ElementType, typename IndptrType, typename IndicesType, typename NZType, bool is_device> class csr_matrix_view : public sparse_matrix_view<ElementType, compressed_structure_view<IndptrType, IndicesType, NZType, is_device>, is_device> { public: using element_type = ElementType; using indptr_type = IndptrType; using indices_type = IndicesType; using nnz_type = NZType; csr_matrix_view( raft::span<ElementType, is_device> element_span, compressed_structure_view<IndptrType, IndicesType, NZType, is_device> structure_view) : sparse_matrix_view<ElementType, compressed_structure_view<IndptrType, IndicesType, NZType, is_device>, is_device>(element_span, structure_view){}; }; template <typename ElementType, typename IndptrType, typename IndicesType, typename NZType, bool is_device, template <typename T> typename ContainerPolicy, SparsityType sparsity_type = SparsityType::OWNING, typename structure_type = std::conditional_t< sparsity_type == SparsityType::OWNING, compressed_structure<IndptrType, IndicesType, NZType, is_device, ContainerPolicy>, compressed_structure_view<IndptrType, IndicesType, NZType, is_device>>> class csr_matrix : public sparse_matrix<ElementType, structure_type, csr_matrix_view<ElementType, IndptrType, IndicesType, NZType, is_device>, is_device, ContainerPolicy> { public: using element_type = ElementType; using indptr_type = IndptrType; using indices_type = IndicesType; using nnz_type = NZType; using structure_view_type = typename structure_type::view_type; static constexpr auto get_sparsity_type() { return sparsity_type; } using sparse_matrix_type = sparse_matrix<ElementType, structure_type, csr_matrix_view<ElementType, IndptrType, IndicesType, NZType, is_device>, is_device, ContainerPolicy>; using container_type = typename ContainerPolicy<ElementType>::container_type; template <SparsityType sparsity_type_ = get_sparsity_type(), typename = typename std::enable_if_t<sparsity_type_ == SparsityType::OWNING>> csr_matrix(raft::resources const& handle, IndptrType n_rows, IndicesType n_cols, NZType nnz = 0) noexcept(std::is_nothrow_default_constructible_v<container_type>) : sparse_matrix_type(handle, n_rows, n_cols, nnz){}; // Constructor that owns the data but not the structure template <SparsityType sparsity_type_ = get_sparsity_type(), typename = typename std::enable_if_t<sparsity_type_ == SparsityType::PRESERVING>> csr_matrix(raft::resources const& handle, structure_type structure) noexcept( std::is_nothrow_default_constructible_v<container_type>) : sparse_matrix_type(handle, structure){}; /** * Initialize the sparsity on this instance if it was not known upon construction * Please note this will resize the underlying memory buffers * @param nnz new sparsity to initialize. */ template <typename = std::enable_if<sparsity_type == SparsityType::OWNING>> void initialize_sparsity(NZType nnz) { sparse_matrix_type::initialize_sparsity(nnz); this->structure_.initialize_sparsity(nnz); } /** * Return a view of the structure underlying this matrix * @return */ structure_view_type structure_view() { if constexpr (get_sparsity_type() == SparsityType::OWNING) { return this->structure_.view(); } else { return this->structure_; } } }; /** @} */ } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/core/mdspan_types.hpp
/* * Copyright (c) 2022-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/thirdparty/mdspan/include/experimental/mdspan> namespace raft { using std::experimental::dynamic_extent; using std::experimental::extents; /** * @defgroup mdspan_layout C- and F-contiguous mdspan layouts * @{ */ using std::experimental::layout_right; using layout_c_contiguous = layout_right; using row_major = layout_right; using std::experimental::layout_left; using layout_f_contiguous = layout_left; using col_major = layout_left; /** @} */ template <typename IndexType> using vector_extent = std::experimental::extents<IndexType, dynamic_extent>; template <typename IndexType> using matrix_extent = std::experimental::extents<IndexType, dynamic_extent, dynamic_extent>; template <typename IndexType> using scalar_extent = std::experimental::extents<IndexType, 1>; /** * @brief Strided layout for non-contiguous memory. */ using std::experimental::layout_stride; template <typename IndexType> using extent_1d = vector_extent<IndexType>; template <typename IndexType> using extent_2d = matrix_extent<IndexType>; template <typename IndexType> using extent_3d = std::experimental::extents<IndexType, dynamic_extent, dynamic_extent, dynamic_extent>; template <typename IndexType> using extent_4d = std::experimental:: extents<IndexType, dynamic_extent, dynamic_extent, dynamic_extent, dynamic_extent>; template <typename IndexType> using extent_5d = std::experimental::extents<IndexType, dynamic_extent, dynamic_extent, dynamic_extent, dynamic_extent, dynamic_extent>; } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/core/interruptible.hpp
/* * Copyright (c) 2021-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #ifndef __RAFT_RT_INTERRUPTIBLE_H #define __RAFT_RT_INTERRUPTIBLE_H #pragma once #include <memory> #include <mutex> #include <optional> #include <raft/core/error.hpp> #include <raft/util/cudart_utils.hpp> #include <rmm/cuda_stream_view.hpp> #include <thread> #include <unordered_map> namespace raft { /** * @defgroup interruptible definitions and classes related to the interruptible API * @{ */ /** * @brief Exception thrown during `interruptible::synchronize` call when it detects a request * to cancel the work performed in this CPU thread. */ struct interrupted_exception : public raft::exception { using raft::exception::exception; }; /** * @brief Cooperative-style interruptible execution. * * This class provides facilities for interrupting execution of a C++ thread at designated points * in code from outside of the thread. In particular, it provides an interruptible version of the * blocking CUDA synchronization function, that allows dropping a long-running GPU work. * * * **Important:** Although CUDA synchronize calls serve as cancellation points, the interruptible * machinery has nothing to do with CUDA streams or events. In other words, when you call `cancel`, * it’s the CPU waiting function what is interrupted, not the GPU stream work. This means, when the * `interrupted_exception` is raised, any unfinished GPU stream work continues to run. It’s the * responsibility of the developer then to make sure the unfinished stream work does not affect the * program in an undesirable way. * * * What can happen to CUDA stream when the `synchronize` is cancelled? If you catch the * `interrupted_exception` immediately, you can safely wait on the stream again. * Otherwise, some of the allocated resources may be released before the active kernel finishes * using them, which will result in writing into deallocated or reallocated memory and undefined * behavior in general. A dead-locked kernel may never finish (or may crash if you’re lucky). In * practice, the outcome is usually acceptable for the use case of emergency program interruption * (e.g., CTRL+C), but extra effort on the use side is required to allow safe interrupting and * resuming of the GPU stream work. */ class interruptible { public: /** * @brief Synchronize the CUDA stream, subject to being interrupted by `interruptible::cancel` * called on this CPU thread. * * @param [in] stream a CUDA stream. * * @throw raft::interrupted_exception if interruptible::cancel() was called on the current CPU * thread before the currently captured work has been finished. * @throw raft::cuda_error if another CUDA error happens. */ static inline void synchronize(rmm::cuda_stream_view stream) { get_token()->synchronize_impl(cudaStreamQuery, stream); } /** * @brief Synchronize the CUDA event, subject to being interrupted by `interruptible::cancel` * called on this CPU thread. * * @param [in] event a CUDA event. * * @throw raft::interrupted_exception if interruptible::cancel() was called on the current CPU * thread before the currently captured work has been finished. * @throw raft::cuda_error if another CUDA error happens. */ static inline void synchronize(cudaEvent_t event) { get_token()->synchronize_impl(cudaEventQuery, event); } /** * @brief Check the thread state, whether the thread can continue execution or is interrupted by * `interruptible::cancel`. * * This is a cancellation point for an interruptible thread. It's called in the internals of * `interruptible::synchronize` in a loop. If two synchronize calls are far apart, it's * recommended to call `interruptible::yield()` in between to make sure the thread does not become * unresponsive for too long. * * Both `yield` and `yield_no_throw` reset the state to non-cancelled after execution. * * @throw raft::interrupted_exception if interruptible::cancel() was called on the current CPU * thread. */ static inline void yield() { get_token()->yield_impl(); } /** * @brief Check the thread state, whether the thread can continue execution or is interrupted by * `interruptible::cancel`. * * Same as `interruptible::yield`, but does not throw an exception if the thread is cancelled. * * Both `yield` and `yield_no_throw` reset the state to non-cancelled after execution. * * @return whether the thread can continue, i.e. `true` means continue, `false` means cancelled. */ static inline auto yield_no_throw() -> bool { return get_token()->yield_no_throw_impl(); } /** * @brief Get a cancellation token for this CPU thread. * * @return an object that can be used to cancel the GPU work waited on this CPU thread. */ static inline auto get_token() -> std::shared_ptr<interruptible> { // NB: using static thread-local storage to keep the token alive once it is initialized static thread_local std::shared_ptr<interruptible> s( get_token_impl<true>(std::this_thread::get_id())); return s; } /** * @brief Get a cancellation token for a CPU thread given by its id. * * The returned token may live longer than the associated thread. In that case, using its * `cancel` method has no effect. * * @param [in] thread_id an id of a C++ CPU thread. * @return an object that can be used to cancel the GPU work waited on the given CPU thread. */ static inline auto get_token(std::thread::id thread_id) -> std::shared_ptr<interruptible> { return get_token_impl<false>(thread_id); } /** * @brief Cancel any current or next call to `interruptible::synchronize` performed on the * CPU thread given by the `thread_id` * * Note, this function uses a mutex to safely get a cancellation token that may be shared * among multiple threads. If you plan to use it from a signal handler, consider the non-static * `cancel()` instead. * * @param [in] thread_id a CPU thread, in which the work should be interrupted. */ static inline void cancel(std::thread::id thread_id) { get_token(thread_id)->cancel(); } /** * @brief Cancel any current or next call to `interruptible::synchronize` performed on the * CPU thread given by this `interruptible` token. * * Note, this function does not involve thread synchronization/locks and does not throw any * exceptions, so it's safe to call from a signal handler. */ inline void cancel() noexcept { continue_.clear(std::memory_order_relaxed); } // don't allow the token to leave the shared_ptr interruptible(interruptible const&) = delete; interruptible(interruptible&&) = delete; auto operator=(interruptible const&) -> interruptible& = delete; auto operator=(interruptible&&) -> interruptible& = delete; private: /** Global registry of thread-local cancellation stores. */ using registry_t = std::tuple<std::mutex, std::unordered_map<std::thread::id, std::weak_ptr<interruptible>>>; /** * The registry "garbage collector": a custom deleter for the interruptible tokens that removes * the token from the registry, if the registry still exists. */ struct registry_gc_t { std::weak_ptr<registry_t> weak_registry; std::thread::id thread_id; inline void operator()(interruptible* thread_store) const noexcept { // the deleter kicks in at thread/program exit; in some cases, the registry_ (static variable) // may have been destructed by this point of time. // Hence, we use a weak pointer to check if the registry still exists. auto registry = weak_registry.lock(); if (registry) { std::lock_guard<std::mutex> guard_erase(std::get<0>(*registry)); auto& map = std::get<1>(*registry); auto found = map.find(thread_id); if (found != map.end()) { auto stored = found->second.lock(); // thread_store is not moveable, thus retains its original location. // Not equal pointers below imply the new store has been already placed // in the registry by the same std::thread::id if (!stored || stored.get() == thread_store) { map.erase(found); } } } delete thread_store; } }; /** * The registry itself is stored in the static memory, in a shared pointer. * This is to safely access it from the destructors of the thread-local tokens. */ static inline std::shared_ptr<registry_t> registry_{new registry_t{}}; /** * Create a new interruptible token or get an existing from the global registry_. * * Presumptions: * * 1. get_token_impl<true> must be called at most once per thread. * 2. When `Claim == true`, thread_id must be equal to std::this_thread::get_id(). * 3. get_token_impl<false> can be called as many times as needed, producing a valid * token for any input thread_id, independent of whether a C++ thread with this * id exists or not. * * @tparam Claim whether to bind the token to the given thread. * @param [in] thread_id the id of the associated C++ thread. * @return new or existing interruptible token. */ template <bool Claim> static auto get_token_impl(std::thread::id thread_id) -> std::shared_ptr<interruptible> { // Make a local copy of the shared pointer to make sure the registry is not destroyed, // if, for any reason, this function is called at program exit. std::shared_ptr<registry_t> shared_registry = registry_; // If the registry is not available, create a lone token that cannot be accessed from // the outside of the thread. if (!shared_registry) { return std::shared_ptr<interruptible>{new interruptible()}; } // Otherwise, proceed with the normal logic std::lock_guard<std::mutex> guard_get(std::get<0>(*shared_registry)); // the following two lines construct an empty shared_ptr if the key does not exist. auto& weak_store = std::get<1>(*shared_registry)[thread_id]; auto thread_store = weak_store.lock(); if (!thread_store || (Claim && thread_store->claimed_)) { // Create a new thread_store in two cases: // 1. It does not exist in the map yet // 2. The previous store in the map has not yet been deleted thread_store.reset(new interruptible(), registry_gc_t{shared_registry, thread_id}); std::weak_ptr<interruptible>(thread_store).swap(weak_store); } // The thread_store is "claimed" by the thread if constexpr (Claim) { thread_store->claimed_ = true; } return thread_store; } /** * Communicate whether the thread is in a cancelled state or can continue execution. * * `yield` checks this flag and always resets it to the signalled state; `cancel` clears it. * These are the only two places where it's used. */ std::atomic_flag continue_; /** This flag is set to true when the created token is placed into a thread-local storage. */ bool claimed_ = false; interruptible() noexcept { yield_no_throw_impl(); } void yield_impl() { if (!yield_no_throw_impl()) { throw interrupted_exception("The work in this thread was cancelled."); } } auto yield_no_throw_impl() noexcept -> bool { return continue_.test_and_set(std::memory_order_relaxed); } template <typename Query, typename Object> inline void synchronize_impl(Query query, Object object) { cudaError_t query_result; while (true) { yield_impl(); query_result = query(object); if (query_result != cudaErrorNotReady) { break; } std::this_thread::yield(); } RAFT_CUDA_TRY(query_result); } }; /** * @} // end doxygen group interruptible */ } // namespace raft #endif
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/core/stream_view.hpp
/* * Copyright (c) 2022-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #include <raft/core/cuda_support.hpp> #include <raft/core/error.hpp> #include <raft/core/logger.hpp> #ifndef RAFT_DISABLE_CUDA #include <raft/core/interruptible.hpp> #include <rmm/cuda_stream_view.hpp> #endif namespace raft { namespace detail { struct fail_stream_view { constexpr fail_stream_view() = default; constexpr fail_stream_view(fail_stream_view const&) = default; constexpr fail_stream_view(fail_stream_view&&) = default; auto constexpr operator=(fail_stream_view const&) -> fail_stream_view& = default; auto constexpr operator=(fail_stream_view&&) -> fail_stream_view& = default; auto value() { throw non_cuda_build_error{"Attempted to access CUDA stream in non-CUDA build"}; } [[nodiscard]] auto is_per_thread_default() const { return false; } [[nodiscard]] auto is_default() const { return false; } void synchronize() const { throw non_cuda_build_error{"Attempted to sync CUDA stream in non-CUDA build"}; } void synchronize_no_throw() const { RAFT_LOG_ERROR("Attempted to sync CUDA stream in non-CUDA build"); } }; } // namespace detail /** A lightweight wrapper around rmm::cuda_stream_view that can be used in * CUDA-free builds * * While CUDA-free builds should never actually make use of a CUDA stream at * runtime, it is sometimes useful to have a symbol that can stand in place of * a CUDA stream to avoid excessive ifdef directives interspersed with other * logic. This struct's methods invoke the underlying rmm::cuda_stream_view in * CUDA-enabled builds but throw runtime exceptions if any non-trivial method * is called from a CUDA-free build */ struct stream_view { #ifndef RAFT_DISABLE_CUDA using underlying_view_type = rmm::cuda_stream_view; #else using underlying_view_type = detail::fail_stream_view; #endif constexpr stream_view( underlying_view_type base_view = stream_view::get_underlying_per_thread_default()) : base_view_{base_view} { } constexpr stream_view(stream_view const&) = default; constexpr stream_view(stream_view&&) = default; auto operator=(stream_view const&) -> stream_view& = default; auto operator=(stream_view&&) -> stream_view& = default; auto value() { return base_view_.value(); } operator underlying_view_type() const noexcept { return base_view_; } [[nodiscard]] auto is_per_thread_default() const { return base_view_.is_per_thread_default(); } [[nodiscard]] auto is_default() const { return base_view_.is_default(); } void synchronize() const { base_view_.synchronize(); } void synchronize_no_throw() const { base_view_.synchronize_no_throw(); } void interruptible_synchronize() const { #ifndef RAFT_DISABLE_CUDA interruptible::synchronize(base_view_); #else synchronize(); #endif } auto underlying() { return base_view_; } void synchronize_if_cuda_enabled() { if constexpr (raft::CUDA_ENABLED) { base_view_.synchronize(); } } private: underlying_view_type base_view_; auto static get_underlying_per_thread_default() -> underlying_view_type { #ifndef RAFT_DISABLE_CUDA return rmm::cuda_stream_per_thread; #else auto static constexpr const default_fail_stream = underlying_view_type{}; return default_fail_stream; #endif } }; auto static const stream_view_per_thread = stream_view{}; } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/core/device_coo_matrix.hpp
/* * Copyright (c) 2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/core/coo_matrix.hpp> #include <raft/core/device_container_policy.hpp> #include <raft/core/device_span.hpp> #include <raft/core/resources.hpp> #include <raft/core/sparse_types.hpp> namespace raft { /** * \defgroup device_coo_matrix Device COO Matrix * @{ */ /** * Specialization for a sparsity-preserving coordinate structure view which uses device memory */ template <typename RowType, typename ColType, typename NZType> using device_coordinate_structure_view = coordinate_structure_view<RowType, ColType, NZType, true>; /** * Specialization for a sparsity-owning coordinate structure which uses device memory */ template <typename RowType, typename ColType, typename NZType, template <typename T> typename ContainerPolicy = device_uvector_policy> using device_coordinate_structure = coordinate_structure<RowType, ColType, NZType, true, ContainerPolicy>; /** * Specialization for a coo matrix view which uses device memory */ template <typename ElementType, typename RowType, typename ColType, typename NZType> using device_coo_matrix_view = coo_matrix_view<ElementType, RowType, ColType, NZType, true>; template <typename ElementType, typename RowType, typename ColType, typename NZType, template <typename T> typename ContainerPolicy = device_uvector_policy, SparsityType sparsity_type = SparsityType::OWNING> using device_coo_matrix = coo_matrix<ElementType, RowType, ColType, NZType, true, ContainerPolicy, sparsity_type>; /** * Specialization for a sparsity-owning coo matrix which uses device memory */ template <typename ElementType, typename RowType, typename ColType, typename NZType, template <typename T> typename ContainerPolicy = device_uvector_policy> using device_sparsity_owning_coo_matrix = coo_matrix<ElementType, RowType, ColType, NZType, true, ContainerPolicy>; template <typename ElementType, typename RowType, typename ColType, typename NZType, template <typename T> typename ContainerPolicy = device_uvector_policy> using device_sparsity_preserving_coo_matrix = coo_matrix<ElementType, RowType, ColType, NZType, true, ContainerPolicy, SparsityType::PRESERVING>; template <typename T> struct is_device_coo_matrix_view : std::false_type {}; template <typename ElementType, typename RowType, typename ColType, typename NZType> struct is_device_coo_matrix_view<device_coo_matrix_view<ElementType, RowType, ColType, NZType>> : std::true_type {}; template <typename T> constexpr bool is_device_coo_matrix_view_v = is_device_coo_matrix_view<T>::value; template <typename T> struct is_device_coo_matrix : std::false_type {}; template <typename ElementType, typename RowType, typename ColType, typename NZType, template <typename T> typename ContainerPolicy, SparsityType sparsity_type> struct is_device_coo_matrix< device_coo_matrix<ElementType, RowType, ColType, NZType, ContainerPolicy, sparsity_type>> : std::true_type {}; template <typename T> constexpr bool is_device_coo_matrix_v = is_device_coo_matrix<T>::value; template <typename T> constexpr bool is_device_coo_sparsity_owning_v = is_device_coo_matrix<T>::value and T::get_sparsity_type() == OWNING; template <typename T> constexpr bool is_device_coo_sparsity_preserving_v = is_device_coo_matrix<T>::value and T::get_sparsity_type() == PRESERVING; /** * Create a sparsity-owning sparse matrix in the coordinate format. sparsity-owning means that * all of the underlying vectors (data, indptr, indices) are owned by the coo_matrix instance. If * not known up front, the sparsity can be ignored in this factory function and `resize()` invoked * on the instance once the sparsity is known. * * @code{.cpp} * #include <raft/core/resources.hpp> * #include <raft/core/device_coo_matrix.hpp> * * int n_rows = 100000; * int n_cols = 10000; * * raft::resources handle; * coo_matrix = raft::make_device_coo_matrix(handle, n_rows, n_cols); * ... * // compute expected sparsity * ... * int nnz = 5000; * coo_matrix.initialize_sparsity(nnz); * @endcode * * @tparam ElementType * @tparam RowType * @tparam ColType * @tparam NZType * @param[in] handle a raft handle for managing expensive device resources * @param[in] n_rows total number of rows in the matrix * @param[in] n_cols total number of columns in the matrix * @param[in] nnz number of non-zeros in the matrix if known [optional] * @return a sparsity-owning sparse matrix in coordinate (coo) format */ template <typename ElementType, typename RowType, typename ColType, typename NZType> auto make_device_coo_matrix(raft::resources const& handle, RowType n_rows, ColType n_cols, NZType nnz = 0) { return device_sparsity_owning_coo_matrix<ElementType, RowType, ColType, NZType>( handle, n_rows, n_cols, nnz); } /** * Create a sparsity-preserving sparse matrix in the coordinate format. sparsity-preserving means * that a view of the coo sparsity is supplied, allowing the values in the sparsity to change but * not the sparsity itself. The csr_matrix instance does not own the sparsity, the sparsity must * be known up front, and cannot be resized later. * * @code{.cpp} * #include <raft/core/resources.hpp> * #include <raft/core/device_coo_matrix.hpp> * * int n_rows = 100000; * int n_cols = 10000; * * raft::resources handle; * coo_structure = raft::make_device_coordinate_structure(handle, n_rows, n_cols); * ... * // compute expected sparsity * ... * coo_structure.initialize_sparsity(nnz); * coo_matrix = raft::make_device_coo_matrix(handle, coo_structure.view()); * @endcode * * @tparam ElementType * @tparam RowType * @tparam ColType * @tparam NZType * @param[in] handle raft handle for managing expensive device resources * @param[in] structure a sparsity-preserving coordinate structural view * @return a sparsity-preserving sparse matrix in coordinate (coo) format */ template <typename ElementType, typename RowType, typename ColType, typename NZType> auto make_device_coo_matrix(raft::resources const& handle, device_coordinate_structure_view<RowType, ColType, NZType> structure) { return device_sparsity_preserving_coo_matrix<ElementType, RowType, ColType, NZType>(handle, structure); } /** * Create a non-owning sparse matrix view in the coordinate format. This is sparsity-preserving, * meaning that the underlying sparsity is known and cannot be changed. Use the sparsity-owning * coo_matrix if sparsity needs to be mutable. * * @code{.cpp} * #include <raft/core/resources.hpp> * #include <raft/core/device_coo_matrix.hpp> * * int n_rows = 100000; * int n_cols = 10000; * int nnz = 5000; * * // The following pointer is assumed to reference device memory for a size of nnz * float* d_elm_ptr = ...; * * raft::resources handle; * coo_structure = raft::make_device_coordinate_structure(handle, n_rows, n_cols, nnz); * coo_matrix_view = raft::make_device_coo_matrix_view(handle, d_elm_ptr, coo_structure.view()); * @endcode * * @tparam ElementType * @tparam RowType * @tparam ColType * @tparam NZType * @param[in] ptr a pointer to array of nonzero matrix elements on device (size nnz) * @param[in] structure a sparsity-preserving coordinate structural view * @return a sparsity-preserving sparse matrix in coordinate (coo) format */ template <typename ElementType, typename RowType, typename ColType, typename NZType> auto make_device_coo_matrix_view( ElementType* ptr, device_coordinate_structure_view<RowType, ColType, NZType> structure) { return device_coo_matrix_view<ElementType, RowType, ColType, NZType>( raft::device_span<ElementType>(ptr, structure.get_nnz()), structure); } /** * Create a non-owning sparse matrix view in the coordinate format. This is sparsity-preserving, * meaning that the underlying sparsity is known and cannot be changed. Use the sparsity-owning * coo_matrix if sparsity needs to be mutable. * * @code{.cpp} * #include <raft/core/resources.hpp> * #include <raft/core/device_span.hpp> * #include <raft/core/device_coo_matrix.hpp> * * int n_rows = 100000; * int n_cols = 10000; * int nnz = 5000; * * // The following span is assumed to be of size nnz * raft::device_span<float> d_elm_ptr; * * raft::resources handle; * coo_structure = raft::make_device_coordinate_structure(handle, n_rows, n_cols, nnz); * coo_matrix_view = raft::make_device_coo_matrix_view(handle, d_elm_ptr, coo_structure.view()); * @endcode * * @tparam ElementType * @tparam RowType * @tparam ColType * @tparam NZType * @param[in] elements a device span containing nonzero matrix elements (size nnz) * @param[in] structure a sparsity-preserving coordinate structural view * @return */ template <typename ElementType, typename RowType, typename ColType, typename NZType> auto make_device_coo_matrix_view( raft::device_span<ElementType> elements, device_coordinate_structure_view<RowType, ColType, NZType> structure) { RAFT_EXPECTS(elements.size() == structure.get_nnz(), "Size of elements must be equal to the nnz from the structure"); return device_coo_matrix_view<ElementType, RowType, ColType, NZType>(elements, structure); } /** * Create a sparsity-owning coordinate structure object. If not known up front, this object can be * resized() once the sparsity (number of non-zeros) is known, postponing the allocation of the * underlying data arrays. * * @code{.cpp} * #include <raft/core/resources.hpp> * #include <raft/core/device_coo_matrix.hpp> * * int n_rows = 100000; * int n_cols = 10000; * int nnz = 5000; * * raft::resources handle; * coo_structure = raft::make_device_coordinate_structure(handle, n_rows, n_cols, nnz); * * ... * // compute expected sparsity * ... * coo_structure.initialize_sparsity(nnz); * @endcode * * @tparam RowType * @tparam ColType * @tparam NZType * @param[in] handle raft handle for managing expensive resources on device * @param[in] n_rows total number of rows * @param[in] n_cols total number of cols * @param[in] nnz number of non-zeros * @return a sparsity-owning coordinate structure instance */ template <typename RowType, typename ColType, typename NZType> auto make_device_coordinate_structure(raft::resources const& handle, RowType n_rows, ColType n_cols, NZType nnz = 0) { return device_coordinate_structure<RowType, ColType, NZType>(handle, n_rows, n_cols, nnz); } /** * Create a non-owning sparsity-preserved coordinate structure view. Sparsity-preserving means that * the underlying sparsity is known and cannot be changed. Use the sparsity-owning version if the * sparsity is not known up front. * * @code{.cpp} * #include <raft/core/resources.hpp> * #include <raft/core/device_coo_matrix.hpp> * * int n_rows = 100000; * int n_cols = 10000; * int nnz = 5000; * * // The following pointers are assumed to reference device memory of size nnz * int *rows = ...; * int *cols = ...; * * raft::resources handle; * coo_structure = raft::make_device_coordinate_structure_view(handle, rows, cols, n_rows, n_cols, * nnz); * @endcode * * @tparam RowType * @tparam ColType * @tparam NZType * @param[in] rows pointer to row indices array on device (size nnz) * @param[in] cols pointer to column indices array on device (size nnz) * @param[in] n_rows total number of rows * @param[in] n_cols total number of columns * @param[in] nnz number of non-zeros * @return a sparsity-preserving coordinate structural view */ template <typename RowType, typename ColType, typename NZType> auto make_device_coordinate_structure_view( RowType* rows, ColType* cols, RowType n_rows, ColType n_cols, NZType nnz) { return device_coordinate_structure_view<RowType, ColType, NZType>( raft::device_span<RowType>(rows, nnz), raft::device_span<ColType>(cols, nnz), n_rows, n_cols); } /** * Create a non-owning sparsity-preserved coordinate structure view. Sparsity-preserving means that * the underlying sparsity is known and cannot be changed. Use the sparsity-owning version if the * sparsity is not known up front. * * @code{.cpp} * #include <raft/core/resources.hpp> * #include <raft/core/device_coo_matrix.hpp> * * int n_rows = 100000; * int n_cols = 10000; * int nnz = 5000; * * // The following device spans are assumed to be of size nnz * raft::device_span<int> rows; * raft::device_span<int> cols; * * raft::resources handle; * coo_structure = raft::make_device_coordinate_structure_view(handle, rows, cols, n_rows, n_cols); * @endcode * * @tparam RowType * @tparam ColType * @tparam NZType * @param[in] rows a device span containing row indices (size nnz) * @param[in] cols a device span containing column indices (size nnz) * @param[in] n_rows total number of rows * @param[in] n_cols total number of columns * @return a sparsity-preserving coordinate structural view */ template <typename RowType, typename ColType, typename NZType> auto make_device_coordinate_structure_view(raft::device_span<RowType> rows, raft::device_span<ColType> cols, RowType n_rows, ColType n_cols) { return device_coordinate_structure_view<RowType, ColType, NZType>(rows, cols, n_rows, n_cols); } /** @} */ }; // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/core/host_device_accessor.hpp
/* * Copyright (c) 2022, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/core/memory_type.hpp> #include <type_traits> namespace raft { /** * @brief A mixin to distinguish host and device memory. This is the primary * accessor used throughout RAFT's APIs to denote whether an underlying pointer * is accessible from device, host, or both. */ template <typename AccessorPolicy, memory_type MemType> struct host_device_accessor : public AccessorPolicy { using accessor_type = AccessorPolicy; auto static constexpr const mem_type = MemType; using is_host_type = std::conditional_t<raft::is_host_accessible(mem_type), std::true_type, std::false_type>; using is_device_type = std::conditional_t<raft::is_device_accessible(mem_type), std::true_type, std::false_type>; using is_managed_type = std::conditional_t<raft::is_host_device_accessible(mem_type), std::true_type, std::false_type>; static constexpr bool is_host_accessible = raft::is_host_accessible(mem_type); static constexpr bool is_device_accessible = raft::is_device_accessible(mem_type); static constexpr bool is_managed_accessible = raft::is_host_device_accessible(mem_type); // make sure the explicit ctor can fall through using AccessorPolicy::AccessorPolicy; using offset_policy = host_device_accessor; host_device_accessor(AccessorPolicy const& that) : AccessorPolicy{that} {} // NOLINT }; } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/core/operators.hpp
/* * Copyright (c) 2022-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <algorithm> #include <cmath> #include <tuple> #include <type_traits> #include <utility> #include <raft/core/detail/macros.hpp> #include <raft/core/math.hpp> namespace raft { /** * @defgroup operators Commonly used functors. * The optional unused arguments are useful for kernels that pass the index along with the value. * @{ */ struct identity_op { template <typename Type, typename... UnusedArgs> constexpr RAFT_INLINE_FUNCTION auto operator()(const Type& in, UnusedArgs...) const { return in; } }; struct void_op { template <typename... UnusedArgs> constexpr RAFT_INLINE_FUNCTION void operator()(UnusedArgs...) const { return; } }; template <typename OutT> struct cast_op { template <typename InT, typename... UnusedArgs> constexpr RAFT_INLINE_FUNCTION auto operator()(InT in, UnusedArgs...) const { return static_cast<OutT>(in); } }; struct key_op { template <typename KVP, typename... UnusedArgs> constexpr RAFT_INLINE_FUNCTION auto operator()(const KVP& p, UnusedArgs...) const { return p.key; } }; struct value_op { template <typename KVP, typename... UnusedArgs> constexpr RAFT_INLINE_FUNCTION auto operator()(const KVP& p, UnusedArgs...) const { return p.value; } }; struct sqrt_op { template <typename Type, typename... UnusedArgs> RAFT_INLINE_FUNCTION auto operator()(const Type& in, UnusedArgs...) const { return raft::sqrt(in); } }; struct nz_op { template <typename Type, typename... UnusedArgs> constexpr RAFT_INLINE_FUNCTION auto operator()(const Type& in, UnusedArgs...) const { return in != Type(0) ? Type(1) : Type(0); } }; struct abs_op { template <typename Type, typename... UnusedArgs> RAFT_INLINE_FUNCTION auto operator()(const Type& in, UnusedArgs...) const { return raft::abs(in); } }; struct sq_op { template <typename Type, typename... UnusedArgs> constexpr RAFT_INLINE_FUNCTION auto operator()(const Type& in, UnusedArgs...) const { return in * in; } }; struct add_op { template <typename T1, typename T2> constexpr RAFT_INLINE_FUNCTION auto operator()(const T1& a, const T2& b) const { return a + b; } }; struct sub_op { template <typename T1, typename T2> constexpr RAFT_INLINE_FUNCTION auto operator()(const T1& a, const T2& b) const { return a - b; } }; struct mul_op { template <typename T1, typename T2> constexpr RAFT_INLINE_FUNCTION auto operator()(const T1& a, const T2& b) const { return a * b; } }; struct div_op { template <typename T1, typename T2> constexpr RAFT_INLINE_FUNCTION auto operator()(const T1& a, const T2& b) const { return a / b; } }; struct div_checkzero_op { template <typename T1, typename T2> constexpr RAFT_INLINE_FUNCTION auto operator()(const T1& a, const T2& b) const { if (b == T2{0}) { return T1{0} / T2{1}; } return a / b; } }; struct pow_op { template <typename Type> RAFT_INLINE_FUNCTION auto operator()(const Type& a, const Type& b) const { return raft::pow(a, b); } }; struct mod_op { template <typename T1, typename T2> constexpr RAFT_INLINE_FUNCTION auto operator()(const T1& a, const T2& b) const { return a % b; } }; struct min_op { template <typename... Args> RAFT_INLINE_FUNCTION auto operator()(Args&&... args) const { return raft::min(std::forward<Args>(args)...); } }; struct max_op { template <typename... Args> RAFT_INLINE_FUNCTION auto operator()(Args&&... args) const { return raft::max(std::forward<Args>(args)...); } }; struct argmin_op { template <typename KVP> constexpr RAFT_INLINE_FUNCTION auto operator()(const KVP& a, const KVP& b) const { if ((b.value < a.value) || ((a.value == b.value) && (b.key < a.key))) { return b; } return a; } }; struct argmax_op { template <typename KVP> constexpr RAFT_INLINE_FUNCTION auto operator()(const KVP& a, const KVP& b) const { if ((b.value > a.value) || ((a.value == b.value) && (b.key < a.key))) { return b; } return a; } }; struct greater_op { template <typename T1, typename T2> constexpr RAFT_INLINE_FUNCTION auto operator()(const T1& a, const T2& b) const { return a > b; } }; struct less_op { template <typename T1, typename T2> constexpr RAFT_INLINE_FUNCTION auto operator()(const T1& a, const T2& b) const { return a < b; } }; struct greater_or_equal_op { template <typename T1, typename T2> constexpr RAFT_INLINE_FUNCTION auto operator()(const T1& a, const T2& b) const { return a >= b; } }; struct less_or_equal_op { template <typename T1, typename T2> constexpr RAFT_INLINE_FUNCTION auto operator()(const T1& a, const T2& b) const { return a <= b; } }; struct equal_op { template <typename T1, typename T2> constexpr RAFT_INLINE_FUNCTION auto operator()(const T1& a, const T2& b) const { return a == b; } }; struct notequal_op { template <typename T1, typename T2> constexpr RAFT_INLINE_FUNCTION auto operator()(const T1& a, const T2& b) const { return a != b; } }; template <typename ScalarT> struct const_op { const ScalarT scalar; constexpr explicit const_op(const ScalarT& s) : scalar{s} {} template <typename... Args> constexpr RAFT_INLINE_FUNCTION auto operator()(Args...) const { return scalar; } }; /** * @brief Wraps around a binary operator, passing a constant on the right-hand side. * * Usage example: * @code{.cpp} * #include <raft/core/operators.hpp> * * raft::plug_const_op<float, raft::mul_op> op(2.0f); * std::cout << op(2.1f) << std::endl; // 4.2 * @endcode * * @tparam ConstT * @tparam BinaryOpT */ template <typename ConstT, typename BinaryOpT> struct plug_const_op { const ConstT c; const BinaryOpT composed_op; template <typename OpT = BinaryOpT, typename UnusedT = std::enable_if_t<std::is_default_constructible_v<OpT>>> constexpr explicit plug_const_op(const ConstT& s) : c{s}, composed_op{} // The compiler complains if composed_op is not initialized explicitly { } constexpr plug_const_op(const ConstT& s, BinaryOpT o) : c{s}, composed_op{o} {} template <typename InT> constexpr RAFT_INLINE_FUNCTION auto operator()(InT a) const { return composed_op(a, c); } }; template <typename Type> using add_const_op = plug_const_op<Type, add_op>; template <typename Type> using sub_const_op = plug_const_op<Type, sub_op>; template <typename Type> using mul_const_op = plug_const_op<Type, mul_op>; template <typename Type> using div_const_op = plug_const_op<Type, div_op>; template <typename Type> using div_checkzero_const_op = plug_const_op<Type, div_checkzero_op>; template <typename Type> using pow_const_op = plug_const_op<Type, pow_op>; template <typename Type> using mod_const_op = plug_const_op<Type, mod_op>; template <typename Type> using mod_const_op = plug_const_op<Type, mod_op>; template <typename Type> using equal_const_op = plug_const_op<Type, equal_op>; /** * @brief Constructs an operator by composing a chain of operators. * * Note that all arguments are passed to the innermost operator. * * Usage example: * @code{.cpp} * #include <raft/core/operators.hpp> * * auto op = raft::compose_op(raft::sqrt_op(), raft::abs_op(), raft::cast_op<float>(), * raft::add_const_op<int>(8)); * std::cout << op(-50) << std::endl; // 6.48074 * @endcode * * @tparam OpsT Any number of operation types. */ template <typename... OpsT> struct compose_op { const std::tuple<OpsT...> ops; template <typename TupleT = std::tuple<OpsT...>, typename CondT = std::enable_if_t<std::is_default_constructible_v<TupleT>>> constexpr compose_op() { } constexpr explicit compose_op(OpsT... ops) : ops{ops...} {} template <typename... Args> constexpr RAFT_INLINE_FUNCTION auto operator()(Args&&... args) const { return compose<sizeof...(OpsT)>(std::forward<Args>(args)...); } private: template <size_t RemOps, typename... Args> constexpr RAFT_INLINE_FUNCTION auto compose(Args&&... args) const { if constexpr (RemOps > 0) { return compose<RemOps - 1>(std::get<RemOps - 1>(ops)(std::forward<Args>(args)...)); } else { return identity_op{}(std::forward<Args>(args)...); } } }; using absdiff_op = compose_op<abs_op, sub_op>; using sqdiff_op = compose_op<sq_op, sub_op>; /** * @brief Constructs an operator by composing an outer op with one inner op for each of its inputs. * * Usage example: * @code{.cpp} * #include <raft/core/operators.hpp> * * raft::map_args_op<raft::add_op, raft::sqrt_op, raft::cast_op<float>> op; * std::cout << op(42.0f, 10) << std::endl; // 16.4807 * @endcode * * @tparam OuterOpT Outer operation type * @tparam ArgOpsT Operation types for each input of the outer operation */ template <typename OuterOpT, typename... ArgOpsT> struct map_args_op { const OuterOpT outer_op; const std::tuple<ArgOpsT...> arg_ops; template <typename T1 = OuterOpT, typename T2 = std::tuple<ArgOpsT...>, typename CondT = std::enable_if_t<std::is_default_constructible_v<T1> && std::is_default_constructible_v<T2>>> constexpr map_args_op() : outer_op{} // The compiler complains if outer_op is not initialized explicitly { } constexpr explicit map_args_op(OuterOpT outer_op, ArgOpsT... arg_ops) : outer_op{outer_op}, arg_ops{arg_ops...} { } template <typename... Args> constexpr RAFT_INLINE_FUNCTION auto operator()(Args&&... args) const { constexpr size_t kNumOps = sizeof...(ArgOpsT); static_assert(kNumOps == sizeof...(Args), "The number of arguments does not match the number of mapping operators"); return map_args(std::make_index_sequence<kNumOps>{}, std::forward<Args>(args)...); } private: template <size_t... I, typename... Args> constexpr RAFT_INLINE_FUNCTION auto map_args(std::index_sequence<I...>, Args&&... args) const { return outer_op(std::get<I>(arg_ops)(std::forward<Args>(args))...); } }; /** @} */ } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/core/logger-inl.hpp
/* * Copyright (c) 2022-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <stdarg.h> #include <algorithm> #include <memory> #include <mutex> #include <sstream> #include <string> #include <unordered_map> #include <stdarg.h> #include "logger-macros.hpp" // The logger-ext.hpp file contains the class declaration of the logger class. // In this case, it is okay to include the logger-ext.hpp file because it // contains no RAFT_EXPLICIT template instantiations. #include "logger-ext.hpp" #define SPDLOG_HEADER_ONLY #include <raft/core/detail/callback_sink.hpp> #include <raft/core/detail/macros.hpp> // RAFT_INLINE_CONDITIONAL #include <spdlog/sinks/stdout_color_sinks.h> // NOLINT #include <spdlog/spdlog.h> // NOLINT namespace raft { namespace detail { inline std::string format(const char* fmt, va_list& vl) { va_list vl_copy; va_copy(vl_copy, vl); int length = std::vsnprintf(nullptr, 0, fmt, vl_copy); assert(length >= 0); std::vector<char> buf(length + 1); std::vsnprintf(buf.data(), length + 1, fmt, vl); return std::string(buf.data()); } RAFT_INLINE_CONDITIONAL std::string format(const char* fmt, ...) { va_list vl; va_start(vl, fmt); std::string str = format(fmt, vl); va_end(vl); return str; } inline int convert_level_to_spdlog(int level) { level = std::max(RAFT_LEVEL_OFF, std::min(RAFT_LEVEL_TRACE, level)); return RAFT_LEVEL_TRACE - level; } } // namespace detail class logger::impl { // defined privately here // ... all private data and functions: all of these // can now change without recompiling callers ... public: std::shared_ptr<spdlog::sinks::callback_sink_mt> sink; std::shared_ptr<spdlog::logger> spdlogger; std::string cur_pattern; int cur_level; impl(std::string const& name_ = "") : sink{std::make_shared<spdlog::sinks::callback_sink_mt>()}, spdlogger{std::make_shared<spdlog::logger>(name_, sink)}, cur_pattern() { } }; // class logger::impl RAFT_INLINE_CONDITIONAL logger::logger(std::string const& name_) : pimpl(new impl(name_)) { set_pattern(default_log_pattern); set_level(RAFT_ACTIVE_LEVEL); } RAFT_INLINE_CONDITIONAL logger& logger::get(std::string const& name) { if (log_map.find(name) == log_map.end()) { log_map[name] = std::make_shared<raft::logger>(name); } return *log_map[name]; } RAFT_INLINE_CONDITIONAL void logger::set_level(int level) { level = raft::detail::convert_level_to_spdlog(level); pimpl->spdlogger->set_level(static_cast<spdlog::level::level_enum>(level)); } RAFT_INLINE_CONDITIONAL void logger::set_pattern(const std::string& pattern) { pimpl->cur_pattern = pattern; pimpl->spdlogger->set_pattern(pattern); } RAFT_INLINE_CONDITIONAL void logger::set_callback(void (*callback)(int lvl, const char* msg)) { pimpl->sink->set_callback(callback); } RAFT_INLINE_CONDITIONAL void logger::set_flush(void (*flush)()) { pimpl->sink->set_flush(flush); } RAFT_INLINE_CONDITIONAL bool logger::should_log_for(int level) const { level = raft::detail::convert_level_to_spdlog(level); auto level_e = static_cast<spdlog::level::level_enum>(level); return pimpl->spdlogger->should_log(level_e); } RAFT_INLINE_CONDITIONAL int logger::get_level() const { auto level_e = pimpl->spdlogger->level(); return RAFT_LEVEL_TRACE - static_cast<int>(level_e); } RAFT_INLINE_CONDITIONAL std::string logger::get_pattern() const { return pimpl->cur_pattern; } RAFT_INLINE_CONDITIONAL void logger::log(int level, const char* fmt, ...) { level = raft::detail::convert_level_to_spdlog(level); auto level_e = static_cast<spdlog::level::level_enum>(level); // explicit check to make sure that we only expand messages when required if (pimpl->spdlogger->should_log(level_e)) { va_list vl; va_start(vl, fmt); auto msg = raft::detail::format(fmt, vl); va_end(vl); pimpl->spdlogger->log(level_e, msg); } } RAFT_INLINE_CONDITIONAL void logger::flush() { pimpl->spdlogger->flush(); } RAFT_INLINE_CONDITIONAL logger::~logger() {} }; // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/core/span.hpp
/* * Copyright (c) 2022-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <cassert> #include <cinttypes> // size_t #include <cstddef> // std::byte #include <raft/core/mdspan_types.hpp> #include <raft/core/detail/macros.hpp> #include <raft/core/detail/span.hpp> // TODO (cjnolet): Remove thrust dependencies here so host_span can be used without CUDA Toolkit // being installed. Reference: https://github.com/rapidsai/raft/issues/812. #include <thrust/distance.h> #include <thrust/functional.h> #include <thrust/host_vector.h> // _RAFT_HOST_DEVICE #include <thrust/iterator/reverse_iterator.h> #include <type_traits> namespace raft { /** * @defgroup span one-dimensional span type * @{ */ /** * @brief The span class defined in ISO C++20. Iterator is defined as plain pointer and * most of the methods have bound check on debug build. * * @code * rmm::device_uvector<float> uvec(10, rmm::cuda_stream_default); * auto view = device_span<float>{uvec.data(), uvec.size()}; * @endcode */ template <typename T, bool is_device, std::size_t Extent = dynamic_extent> class span { public: using element_type = T; using value_type = typename std::remove_cv<T>::type; using size_type = std::size_t; using difference_type = std::ptrdiff_t; using pointer = T*; using const_pointer = T const*; using reference = T&; using const_reference = T const&; using iterator = pointer; using const_iterator = const_pointer; using reverse_iterator = thrust::reverse_iterator<iterator>; using const_reverse_iterator = thrust::reverse_iterator<const_iterator>; /** * @brief Default constructor that constructs a span with size 0 and nullptr. */ constexpr span() noexcept = default; /** * @brief Constructs a span that is a view over the range [first, first + count); */ constexpr span(pointer ptr, size_type count) noexcept : storage_{ptr, count} { assert(!(Extent != dynamic_extent && count != Extent)); assert(ptr || count == 0); } /** * @brief Constructs a span that is a view over the range [first, last) */ constexpr span(pointer first, pointer last) noexcept : span{first, static_cast<size_type>(thrust::distance(first, last))} { } /** * @brief Constructs a span that is a view over the array arr. */ template <std::size_t N> constexpr span(element_type (&arr)[N]) noexcept : span{&arr[0], N} { } /** * @brief Initialize a span class from another one who's underlying type is convertible * to element_type. */ template <class U, std::size_t OtherExtent, class = typename std::enable_if< detail::is_allowed_element_type_conversion_t<U, T>::value && detail::is_allowed_extent_conversion_t<OtherExtent, Extent>::value>> constexpr span(const span<U, is_device, OtherExtent>& other) noexcept : span{other.data(), other.size()} { } constexpr span(span const& other) noexcept = default; constexpr span(span&& other) noexcept = default; constexpr auto operator=(span const& other) noexcept -> span& = default; constexpr auto operator=(span&& other) noexcept -> span& = default; constexpr auto begin() const noexcept -> iterator { return data(); } constexpr auto end() const noexcept -> iterator { return data() + size(); } constexpr auto cbegin() const noexcept -> const_iterator { return data(); } constexpr auto cend() const noexcept -> const_iterator { return data() + size(); } _RAFT_HOST_DEVICE constexpr auto rbegin() const noexcept -> reverse_iterator { return reverse_iterator{end()}; } _RAFT_HOST_DEVICE constexpr auto rend() const noexcept -> reverse_iterator { return reverse_iterator{begin()}; } _RAFT_HOST_DEVICE constexpr auto crbegin() const noexcept -> const_reverse_iterator { return const_reverse_iterator{cend()}; } _RAFT_HOST_DEVICE constexpr auto crend() const noexcept -> const_reverse_iterator { return const_reverse_iterator{cbegin()}; } // element access constexpr auto front() const -> reference { return (*this)[0]; } constexpr auto back() const -> reference { return (*this)[size() - 1]; } template <typename Index> constexpr auto operator[](Index _idx) const -> reference { assert(static_cast<size_type>(_idx) < size()); return data()[_idx]; } constexpr auto data() const noexcept -> pointer { return storage_.data(); } // Observers [[nodiscard]] constexpr auto size() const noexcept -> size_type { return storage_.size(); } [[nodiscard]] constexpr auto size_bytes() const noexcept -> size_type { return size() * sizeof(T); } constexpr auto empty() const noexcept { return size() == 0; } // Subviews template <std::size_t Count> constexpr auto first() const -> span<element_type, is_device, Count> { assert(Count <= size()); return {data(), Count}; } constexpr auto first(std::size_t _count) const -> span<element_type, is_device, dynamic_extent> { assert(_count <= size()); return {data(), _count}; } template <std::size_t Count> constexpr auto last() const -> span<element_type, is_device, Count> { assert(Count <= size()); return {data() + size() - Count, Count}; } constexpr auto last(std::size_t _count) const -> span<element_type, is_device, dynamic_extent> { assert(_count <= size()); return subspan(size() - _count, _count); } /*! * If Count is std::dynamic_extent, r.size() == this->size() - Offset; * Otherwise r.size() == Count. */ template <std::size_t Offset, std::size_t Count = dynamic_extent> constexpr auto subspan() const -> span<element_type, is_device, detail::extent_value_t<Extent, Offset, Count>::value> { assert((Count == dynamic_extent) ? (Offset <= size()) : (Offset + Count <= size())); return {data() + Offset, Count == dynamic_extent ? size() - Offset : Count}; } constexpr auto subspan(size_type _offset, size_type _count = dynamic_extent) const -> span<element_type, is_device, dynamic_extent> { assert((_count == dynamic_extent) ? (_offset <= size()) : (_offset + _count <= size())); return {data() + _offset, _count == dynamic_extent ? size() - _offset : _count}; } private: detail::span_storage<T, Extent> storage_; }; template <class T, std::size_t X, class U, std::size_t Y, bool is_device> constexpr auto operator==(span<T, is_device, X> l, span<U, is_device, Y> r) -> bool { if (l.size() != r.size()) { return false; } for (auto l_beg = l.cbegin(), r_beg = r.cbegin(); l_beg != l.cend(); ++l_beg, ++r_beg) { if (*l_beg != *r_beg) { return false; } } return true; } template <class T, std::size_t X, class U, std::size_t Y, bool is_device> constexpr auto operator!=(span<T, is_device, X> l, span<U, is_device, Y> r) { return !(l == r); } template <class T, std::size_t X, class U, std::size_t Y, bool is_device> constexpr auto operator<(span<T, is_device, X> l, span<U, is_device, Y> r) { return detail::lexicographical_compare< typename span<T, is_device, X>::iterator, typename span<U, is_device, Y>::iterator, thrust::less<typename span<T, is_device, X>::element_type>>( l.begin(), l.end(), r.begin(), r.end()); } template <class T, std::size_t X, class U, std::size_t Y, bool is_device> constexpr auto operator<=(span<T, is_device, X> l, span<U, is_device, Y> r) { return !(l > r); } template <class T, std::size_t X, class U, std::size_t Y, bool is_device> constexpr auto operator>(span<T, is_device, X> l, span<U, is_device, Y> r) { return detail::lexicographical_compare< typename span<T, is_device, X>::iterator, typename span<U, is_device, Y>::iterator, thrust::greater<typename span<T, is_device, X>::element_type>>( l.begin(), l.end(), r.begin(), r.end()); } template <class T, std::size_t X, class U, std::size_t Y, bool is_device> constexpr auto operator>=(span<T, is_device, X> l, span<U, is_device, Y> r) { return !(l < r); } /** * @brief Converts a span into a view of its underlying bytes */ template <class T, bool is_device, std::size_t E> auto as_bytes(span<T, is_device, E> s) noexcept -> span<const std::byte, is_device, detail::extent_as_bytes_value_t<T, E>::value> { return {reinterpret_cast<const std::byte*>(s.data()), s.size_bytes()}; } /** * @brief Converts a span into a mutable view of its underlying bytes */ template <class T, bool is_device, std::size_t E> auto as_writable_bytes(span<T, is_device, E> s) noexcept -> span<std::byte, is_device, detail::extent_as_bytes_value_t<T, E>::value> { return {reinterpret_cast<std::byte*>(s.data()), s.size_bytes()}; } /* @} */ } // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/core/cublas_macros.hpp
/* * Copyright (c) 2022-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #ifndef __RAFT_RT_CUBLAS_MACROS_H #define __RAFT_RT_CUBLAS_MACROS_H #pragma once #include <cublas_v2.h> #include <raft/core/error.hpp> ///@todo: enable this once we have logger enabled // #include <cuml/common/logger.hpp> #include <cstdint> #define _CUBLAS_ERR_TO_STR(err) \ case err: return #err namespace raft { /** * @ingroup error_handling * @{ */ /** * @brief Exception thrown when a cuBLAS error is encountered. */ struct cublas_error : public raft::exception { explicit cublas_error(char const* const message) : raft::exception(message) {} explicit cublas_error(std::string const& message) : raft::exception(message) {} }; /** * @} */ namespace linalg { namespace detail { inline const char* cublas_error_to_string(cublasStatus_t err) { switch (err) { _CUBLAS_ERR_TO_STR(CUBLAS_STATUS_SUCCESS); _CUBLAS_ERR_TO_STR(CUBLAS_STATUS_NOT_INITIALIZED); _CUBLAS_ERR_TO_STR(CUBLAS_STATUS_ALLOC_FAILED); _CUBLAS_ERR_TO_STR(CUBLAS_STATUS_INVALID_VALUE); _CUBLAS_ERR_TO_STR(CUBLAS_STATUS_ARCH_MISMATCH); _CUBLAS_ERR_TO_STR(CUBLAS_STATUS_MAPPING_ERROR); _CUBLAS_ERR_TO_STR(CUBLAS_STATUS_EXECUTION_FAILED); _CUBLAS_ERR_TO_STR(CUBLAS_STATUS_INTERNAL_ERROR); _CUBLAS_ERR_TO_STR(CUBLAS_STATUS_NOT_SUPPORTED); _CUBLAS_ERR_TO_STR(CUBLAS_STATUS_LICENSE_ERROR); default: return "CUBLAS_STATUS_UNKNOWN"; }; } } // namespace detail } // namespace linalg } // namespace raft #undef _CUBLAS_ERR_TO_STR /** * @ingroup assertion * @{ */ /** * @brief Error checking macro for cuBLAS runtime API functions. * * Invokes a cuBLAS runtime API function call, if the call does not return * CUBLAS_STATUS_SUCCESS, throws an exception detailing the cuBLAS error that occurred */ #define RAFT_CUBLAS_TRY(call) \ do { \ cublasStatus_t const status = (call); \ if (CUBLAS_STATUS_SUCCESS != status) { \ std::string msg{}; \ SET_ERROR_MSG(msg, \ "cuBLAS error encountered at: ", \ "call='%s', Reason=%d:%s", \ #call, \ status, \ raft::linalg::detail::cublas_error_to_string(status)); \ throw raft::cublas_error(msg); \ } \ } while (0) // FIXME: Remove after consumers rename #ifndef CUBLAS_TRY #define CUBLAS_TRY(call) RAFT_CUBLAS_TRY(call) #endif // /** // * @brief check for cuda runtime API errors but log error instead of raising // * exception. // */ #define RAFT_CUBLAS_TRY_NO_THROW(call) \ do { \ cublasStatus_t const status = call; \ if (CUBLAS_STATUS_SUCCESS != status) { \ printf("CUBLAS call='%s' at file=%s line=%d failed with %s\n", \ #call, \ __FILE__, \ __LINE__, \ raft::linalg::detail::cublas_error_to_string(status)); \ } \ } while (0) /** * @} */ /** FIXME: remove after cuml rename */ #ifndef CUBLAS_CHECK #define CUBLAS_CHECK(call) CUBLAS_TRY(call) #endif /** FIXME: remove after cuml rename */ #ifndef CUBLAS_CHECK_NO_THROW #define CUBLAS_CHECK_NO_THROW(call) RAFT_CUBLAS_TRY_NO_THROW(call) #endif #endif
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/core/logger-ext.hpp
/* * Copyright (c) 2022-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <memory> // std::unique_ptr #include <raft/core/detail/macros.hpp> // RAFT_INLINE_CONDITIONAL #include <string> // std::string #include <unordered_map> // std::unordered_map namespace raft { static const std::string RAFT_NAME = "raft"; static const std::string default_log_pattern("[%L] [%H:%M:%S.%f] %v"); namespace detail { RAFT_INLINE_CONDITIONAL std::string format(const char* fmt, ...); } /** * @brief The main Logging class for raft library. * * This class acts as a thin wrapper over the underlying `spdlog` interface. The * design is done in this way in order to avoid us having to also ship `spdlog` * header files in our installation. * * @todo This currently only supports logging to stdout. Need to add support in * future to add custom loggers as well [Issue #2046] */ class logger { public: // @todo setting the logger once per process with logger(std::string const& name_ = ""); /** * @brief Singleton method to get the underlying logger object * * @return the singleton logger object */ static logger& get(std::string const& name = ""); /** * @brief Set the logging level. * * Only messages with level equal or above this will be printed * * @param[in] level logging level * * @note The log level will actually be set only if the input is within the * range [RAFT_LEVEL_TRACE, RAFT_LEVEL_OFF]. If it is not, then it'll * be ignored. See documentation of decisiontree for how this gets used */ void set_level(int level); /** * @brief Set the logging pattern * * @param[in] pattern the pattern to be set. Refer this link * https://github.com/gabime/spdlog/wiki/3.-Custom-formatting * to know the right syntax of this pattern */ void set_pattern(const std::string& pattern); /** * @brief Register a callback function to be run in place of usual log call * * @param[in] callback the function to be run on all logged messages */ void set_callback(void (*callback)(int lvl, const char* msg)); /** * @brief Register a flush function compatible with the registered callback * * @param[in] flush the function to use when flushing logs */ void set_flush(void (*flush)()); /** * @brief Tells whether messages will be logged for the given log level * * @param[in] level log level to be checked for * @return true if messages will be logged for this level, else false */ bool should_log_for(int level) const; /** * @brief Query for the current log level * * @return the current log level */ int get_level() const; /** * @brief Get the current logging pattern * @return the pattern */ std::string get_pattern() const; /** * @brief Main logging method * * @param[in] level logging level of this message * @param[in] fmt C-like format string, followed by respective params */ void log(int level, const char* fmt, ...); /** * @brief Flush logs by calling flush on underlying logger */ void flush(); ~logger(); private: logger(); // pimpl pattern: // https://learn.microsoft.com/en-us/cpp/cpp/pimpl-for-compile-time-encapsulation-modern-cpp?view=msvc-170 class impl; std::unique_ptr<impl> pimpl; static inline std::unordered_map<std::string, std::shared_ptr<raft::logger>> log_map; }; // class logger /** * @brief An object used for scoped log level setting * * Instances of `raft::log_level_setter` will set RAFT logging to the level * indicated on construction and will revert to the previous set level on * destruction. */ struct log_level_setter { explicit log_level_setter(int level) { prev_level_ = logger::get(RAFT_NAME).get_level(); logger::get(RAFT_NAME).set_level(level); } ~log_level_setter() { logger::get(RAFT_NAME).set_level(prev_level_); } private: int prev_level_; }; // class log_level_setter }; // namespace raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/core/handle.hpp
/* * Copyright (c) 2019-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #pragma once #include <raft/core/device_resources.hpp> namespace raft { /** * raft::handle_t is being kept around for backwards * compatibility and will be removed in a future version. * * Extending the `raft::handle_t` instead of `using` to * minimize needed changes downstream * (e.g. existing forward declarations, etc...) * * Use of `raft::resources` or `raft::handle_t` is preferred. */ class handle_t : public raft::device_resources { public: handle_t(const handle_t& handle, std::shared_ptr<rmm::mr::device_memory_resource> workspace_resource) : device_resources(handle, workspace_resource) { } handle_t(const handle_t& handle) : device_resources{handle} {} handle_t(handle_t&&) = delete; handle_t& operator=(handle_t&&) = delete; /** * @brief Construct a resources instance with a stream view and stream pool * * @param[in] stream_view the default stream (which has the default per-thread stream if * unspecified) * @param[in] stream_pool the stream pool used (which has default of nullptr if unspecified) * @param[in] workspace_resource an optional resource used by some functions for allocating * temporary workspaces. */ handle_t(rmm::cuda_stream_view stream_view = rmm::cuda_stream_per_thread, std::shared_ptr<rmm::cuda_stream_pool> stream_pool = {nullptr}, std::shared_ptr<rmm::mr::device_memory_resource> workspace_resource = {nullptr}) : device_resources{stream_view, stream_pool, workspace_resource} { } /** Destroys all held-up resources */ ~handle_t() override {} }; } // end NAMESPACE raft
0
rapidsai_public_repos/raft/cpp/include/raft
rapidsai_public_repos/raft/cpp/include/raft/core/device_resources.hpp
/* * Copyright (c) 2019-2023, NVIDIA CORPORATION. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ #ifndef __RAFT_DEVICE_RESOURCES #define __RAFT_DEVICE_RESOURCES #pragma once #include <memory> #include <mutex> #include <optional> #include <string> #include <unordered_map> #include <utility> #include <vector> #include <cublas_v2.h> #include <cuda_runtime.h> #include <cusolverDn.h> #include <cusolverSp.h> #include <cusparse.h> #include <raft/core/comms.hpp> #include <rmm/cuda_stream_pool.hpp> #include <rmm/exec_policy.hpp> #include <raft/core/resource/comms.hpp> #include <raft/core/resource/cublas_handle.hpp> #include <raft/core/resource/cuda_event.hpp> #include <raft/core/resource/cuda_stream.hpp> #include <raft/core/resource/cuda_stream_pool.hpp> #include <raft/core/resource/cusolver_dn_handle.hpp> #include <raft/core/resource/cusolver_sp_handle.hpp> #include <raft/core/resource/cusparse_handle.hpp> #include <raft/core/resource/device_id.hpp> #include <raft/core/resource/device_memory_resource.hpp> #include <raft/core/resource/device_properties.hpp> #include <raft/core/resource/sub_comms.hpp> #include <raft/core/resource/thrust_policy.hpp> #include <raft/core/resources.hpp> namespace raft { /** * @brief Main resource container object that stores all necessary resources * used for calling necessary device functions, cuda kernels and/or libraries */ class device_resources : public resources { public: device_resources(const device_resources& handle, std::shared_ptr<rmm::mr::device_memory_resource> workspace_resource, std::optional<std::size_t> allocation_limit = std::nullopt) : resources{handle} { // replace the resource factory for the workspace_resources resource::set_workspace_resource(*this, workspace_resource, allocation_limit); } device_resources(const device_resources& handle) : resources{handle} {} device_resources(device_resources&&) = delete; device_resources& operator=(device_resources&&) = delete; /** * @brief Construct a resources instance with a stream view and stream pool * * @param[in] stream_view the default stream (which has the default per-thread stream if * unspecified) * @param[in] stream_pool the stream pool used (which has default of nullptr if unspecified) * @param[in] workspace_resource an optional resource used by some functions for allocating * temporary workspaces. * @param[in] allocation_limit the total amount of memory in bytes available to the temporary * workspace resources. */ device_resources(rmm::cuda_stream_view stream_view = rmm::cuda_stream_per_thread, std::shared_ptr<rmm::cuda_stream_pool> stream_pool = {nullptr}, std::shared_ptr<rmm::mr::device_memory_resource> workspace_resource = {nullptr}, std::optional<std::size_t> allocation_limit = std::nullopt) : resources{} { resources::add_resource_factory(std::make_shared<resource::device_id_resource_factory>()); resources::add_resource_factory( std::make_shared<resource::cuda_stream_resource_factory>(stream_view)); resources::add_resource_factory( std::make_shared<resource::cuda_stream_pool_resource_factory>(stream_pool)); if (workspace_resource) { resource::set_workspace_resource(*this, workspace_resource, allocation_limit); } } /** Destroys all held-up resources */ virtual ~device_resources() {} int get_device() const { return resource::get_device_id(*this); } cublasHandle_t get_cublas_handle() const { return resource::get_cublas_handle(*this); } cusolverDnHandle_t get_cusolver_dn_handle() const { return resource::get_cusolver_dn_handle(*this); } cusolverSpHandle_t get_cusolver_sp_handle() const { return resource::get_cusolver_sp_handle(*this); } cusparseHandle_t get_cusparse_handle() const { return resource::get_cusparse_handle(*this); } rmm::exec_policy& get_thrust_policy() const { return resource::get_thrust_policy(*this); } /** * @brief synchronize a stream on the current container */ void sync_stream(rmm::cuda_stream_view stream) const { resource::sync_stream(*this, stream); } /** * @brief synchronize main stream on the current container */ void sync_stream() const { resource::sync_stream(*this); } /** * @brief returns main stream on the current container */ rmm::cuda_stream_view get_stream() const { return resource::get_cuda_stream(*this); } /** * @brief returns whether stream pool was initialized on the current container */ bool is_stream_pool_initialized() const { return resource::is_stream_pool_initialized(*this); } /** * @brief returns stream pool on the current container */ const rmm::cuda_stream_pool& get_stream_pool() const { return resource::get_cuda_stream_pool(*this); } std::size_t get_stream_pool_size() const { return resource::get_stream_pool_size(*this); } /** * @brief return stream from pool */ rmm::cuda_stream_view get_stream_from_stream_pool() const { return resource::get_stream_from_stream_pool(*this); } /** * @brief return stream from pool at index */ rmm::cuda_stream_view get_stream_from_stream_pool(std::size_t stream_idx) const { return resource::get_stream_from_stream_pool(*this, stream_idx); } /** * @brief return stream from pool if size > 0, else main stream on current container */ rmm::cuda_stream_view get_next_usable_stream() const { return resource::get_next_usable_stream(*this); } /** * @brief return stream from pool at index if size > 0, else main stream on current container * * @param[in] stream_idx the required index of the stream in the stream pool if available */ rmm::cuda_stream_view get_next_usable_stream(std::size_t stream_idx) const { return resource::get_next_usable_stream(*this, stream_idx); } /** * @brief synchronize the stream pool on the current container */ void sync_stream_pool() const { return resource::sync_stream_pool(*this); } /** * @brief synchronize subset of stream pool * * @param[in] stream_indices the indices of the streams in the stream pool to synchronize */ void sync_stream_pool(const std::vector<std::size_t> stream_indices) const { return resource::sync_stream_pool(*this, stream_indices); } /** * @brief ask stream pool to wait on last event in main stream */ void wait_stream_pool_on_stream() const { return resource::wait_stream_pool_on_stream(*this); } void set_comms(std::shared_ptr<comms::comms_t> communicator) { resource::set_comms(*this, communicator); } const comms::comms_t& get_comms() const { return resource::get_comms(*this); } void set_subcomm(std::string key, std::shared_ptr<comms::comms_t> subcomm) { resource::set_subcomm(*this, key, subcomm); } const comms::comms_t& get_subcomm(std::string key) const { return resource::get_subcomm(*this, key); } rmm::mr::device_memory_resource* get_workspace_resource() const { return resource::get_workspace_resource(*this); } bool comms_initialized() const { return resource::comms_initialized(*this); } const cudaDeviceProp& get_device_properties() const { return resource::get_device_properties(*this); } }; // class device_resources /** * @brief RAII approach to synchronizing across all streams in the current container */ class stream_syncer { public: explicit stream_syncer(const device_resources& handle) : handle_(handle) { resource::sync_stream(handle_); } ~stream_syncer() { handle_.wait_stream_pool_on_stream(); handle_.sync_stream_pool(); } stream_syncer(const stream_syncer& other) = delete; stream_syncer& operator=(const stream_syncer& other) = delete; private: const device_resources& handle_; }; // class stream_syncer } // namespace raft #endif
0