Add files using upload-large-folder tool
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- .gitattributes +2 -0
- .venv/lib/python3.11/site-packages/numpy/core/tests/__pycache__/test_nditer.cpython-311.pyc +3 -0
- .venv/lib/python3.11/site-packages/numpy/core/tests/__pycache__/test_numeric.cpython-311.pyc +3 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/_shell_utils.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/armccompiler.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/ccompiler.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/conv_template.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/cpuinfo.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/exec_command.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/extension.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/from_template.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/fujitsuccompiler.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/intelccompiler.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/msvc9compiler.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/msvccompiler.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/npy_pkg_config.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/pathccompiler.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/unixccompiler.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_asimddp.c +16 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_avx.c +20 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_avx512_clx.c +22 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_avx512_knl.c +25 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_avx512_knm.c +30 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_avx512_skx.c +26 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_avx512_spr.c +26 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_avx512cd.c +20 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_avx512f.c +20 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_fma4.c +13 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_neon_fp16.c +11 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_neon_vfpv4.c +21 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_popcnt.c +32 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_sse.c +20 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_sse2.c +20 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_sse3.c +20 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_sse41.c +20 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_ssse3.c +20 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_vsx3.c +13 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_vx.c +16 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_vxe.c +25 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_xop.c +12 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/extra_avx512bw_mask.c +18 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/checks/extra_avx512dq_mask.c +16 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/command/__init__.py +41 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/command/__pycache__/__init__.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/command/__pycache__/autodist.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/command/__pycache__/bdist_rpm.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/command/__pycache__/build.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/command/__pycache__/build_clib.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/command/__pycache__/build_ext.cpython-311.pyc +0 -0
- .venv/lib/python3.11/site-packages/numpy/distutils/command/__pycache__/build_py.cpython-311.pyc +0 -0
.gitattributes
CHANGED
|
@@ -380,3 +380,5 @@ tuning-competition-baseline/.venv/lib/python3.11/site-packages/nvidia/cudnn/lib/
|
|
| 380 |
.venv/lib/python3.11/site-packages/numpy/core/tests/__pycache__/test_datetime.cpython-311.pyc filter=lfs diff=lfs merge=lfs -text
|
| 381 |
.venv/lib/python3.11/site-packages/numpy/core/tests/__pycache__/test_umath.cpython-311.pyc filter=lfs diff=lfs merge=lfs -text
|
| 382 |
.venv/lib/python3.11/site-packages/numpy/core/tests/__pycache__/test_ufunc.cpython-311.pyc filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
| 380 |
.venv/lib/python3.11/site-packages/numpy/core/tests/__pycache__/test_datetime.cpython-311.pyc filter=lfs diff=lfs merge=lfs -text
|
| 381 |
.venv/lib/python3.11/site-packages/numpy/core/tests/__pycache__/test_umath.cpython-311.pyc filter=lfs diff=lfs merge=lfs -text
|
| 382 |
.venv/lib/python3.11/site-packages/numpy/core/tests/__pycache__/test_ufunc.cpython-311.pyc filter=lfs diff=lfs merge=lfs -text
|
| 383 |
+
.venv/lib/python3.11/site-packages/numpy/core/tests/__pycache__/test_nditer.cpython-311.pyc filter=lfs diff=lfs merge=lfs -text
|
| 384 |
+
.venv/lib/python3.11/site-packages/numpy/core/tests/__pycache__/test_numeric.cpython-311.pyc filter=lfs diff=lfs merge=lfs -text
|
.venv/lib/python3.11/site-packages/numpy/core/tests/__pycache__/test_nditer.cpython-311.pyc
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c77a1edab8f9c05fe34cfa70cbff3e22696c94ccb808775c331e54e3df71cff2
|
| 3 |
+
size 214681
|
.venv/lib/python3.11/site-packages/numpy/core/tests/__pycache__/test_numeric.cpython-311.pyc
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7489bb86c7b5b58b702dd3b5fefba64fc3720c1a14b2224634f952bb8cb81ea4
|
| 3 |
+
size 275370
|
.venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/_shell_utils.cpython-311.pyc
ADDED
|
Binary file (4.75 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/armccompiler.cpython-311.pyc
ADDED
|
Binary file (1.28 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/ccompiler.cpython-311.pyc
ADDED
|
Binary file (34.6 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/conv_template.cpython-311.pyc
ADDED
|
Binary file (14.2 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/cpuinfo.cpython-311.pyc
ADDED
|
Binary file (50.2 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/exec_command.cpython-311.pyc
ADDED
|
Binary file (15 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/extension.cpython-311.pyc
ADDED
|
Binary file (3.62 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/from_template.cpython-311.pyc
ADDED
|
Binary file (12.8 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/fujitsuccompiler.cpython-311.pyc
ADDED
|
Binary file (1.32 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/intelccompiler.cpython-311.pyc
ADDED
|
Binary file (5.58 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/msvc9compiler.cpython-311.pyc
ADDED
|
Binary file (3.11 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/msvccompiler.cpython-311.pyc
ADDED
|
Binary file (3.39 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/npy_pkg_config.cpython-311.pyc
ADDED
|
Binary file (19.9 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/pathccompiler.cpython-311.pyc
ADDED
|
Binary file (1.26 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/__pycache__/unixccompiler.cpython-311.pyc
ADDED
|
Binary file (6.33 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_asimddp.c
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#ifdef _MSC_VER
|
| 2 |
+
#include <Intrin.h>
|
| 3 |
+
#endif
|
| 4 |
+
#include <arm_neon.h>
|
| 5 |
+
|
| 6 |
+
int main(int argc, char **argv)
|
| 7 |
+
{
|
| 8 |
+
unsigned char *src = (unsigned char*)argv[argc-1];
|
| 9 |
+
uint8x16_t v1 = vdupq_n_u8(src[0]), v2 = vdupq_n_u8(src[1]);
|
| 10 |
+
uint32x4_t va = vdupq_n_u32(3);
|
| 11 |
+
int ret = (int)vgetq_lane_u32(vdotq_u32(va, v1, v2), 0);
|
| 12 |
+
#ifdef __aarch64__
|
| 13 |
+
ret += (int)vgetq_lane_u32(vdotq_laneq_u32(va, v1, v2, 0), 0);
|
| 14 |
+
#endif
|
| 15 |
+
return ret;
|
| 16 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_avx.c
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#if defined(DETECT_FEATURES) && defined(__INTEL_COMPILER)
|
| 2 |
+
/*
|
| 3 |
+
* Unlike GCC and CLANG, Intel Compiler exposes all supported intrinsics,
|
| 4 |
+
* whether or not the build options for those features are specified.
|
| 5 |
+
* Therefore, we must test #definitions of CPU features when option native/host
|
| 6 |
+
* is enabled via `--cpu-baseline` or through env var `CFLAGS` otherwise
|
| 7 |
+
* the test will be broken and leads to enable all possible features.
|
| 8 |
+
*/
|
| 9 |
+
#ifndef __AVX__
|
| 10 |
+
#error "HOST/ARCH doesn't support AVX"
|
| 11 |
+
#endif
|
| 12 |
+
#endif
|
| 13 |
+
|
| 14 |
+
#include <immintrin.h>
|
| 15 |
+
|
| 16 |
+
int main(int argc, char **argv)
|
| 17 |
+
{
|
| 18 |
+
__m256 a = _mm256_add_ps(_mm256_loadu_ps((const float*)argv[argc-1]), _mm256_loadu_ps((const float*)argv[1]));
|
| 19 |
+
return (int)_mm_cvtss_f32(_mm256_castps256_ps128(a));
|
| 20 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_avx512_clx.c
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#if defined(DETECT_FEATURES) && defined(__INTEL_COMPILER)
|
| 2 |
+
/*
|
| 3 |
+
* Unlike GCC and CLANG, Intel Compiler exposes all supported intrinsics,
|
| 4 |
+
* whether or not the build options for those features are specified.
|
| 5 |
+
* Therefore, we must test #definitions of CPU features when option native/host
|
| 6 |
+
* is enabled via `--cpu-baseline` or through env var `CFLAGS` otherwise
|
| 7 |
+
* the test will be broken and leads to enable all possible features.
|
| 8 |
+
*/
|
| 9 |
+
#ifndef __AVX512VNNI__
|
| 10 |
+
#error "HOST/ARCH doesn't support CascadeLake AVX512 features"
|
| 11 |
+
#endif
|
| 12 |
+
#endif
|
| 13 |
+
|
| 14 |
+
#include <immintrin.h>
|
| 15 |
+
|
| 16 |
+
int main(int argc, char **argv)
|
| 17 |
+
{
|
| 18 |
+
/* VNNI */
|
| 19 |
+
__m512i a = _mm512_loadu_si512((const __m512i*)argv[argc-1]);
|
| 20 |
+
a = _mm512_dpbusd_epi32(a, _mm512_setzero_si512(), a);
|
| 21 |
+
return _mm_cvtsi128_si32(_mm512_castsi512_si128(a));
|
| 22 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_avx512_knl.c
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#if defined(DETECT_FEATURES) && defined(__INTEL_COMPILER)
|
| 2 |
+
/*
|
| 3 |
+
* Unlike GCC and CLANG, Intel Compiler exposes all supported intrinsics,
|
| 4 |
+
* whether or not the build options for those features are specified.
|
| 5 |
+
* Therefore, we must test #definitions of CPU features when option native/host
|
| 6 |
+
* is enabled via `--cpu-baseline` or through env var `CFLAGS` otherwise
|
| 7 |
+
* the test will be broken and leads to enable all possible features.
|
| 8 |
+
*/
|
| 9 |
+
#if !defined(__AVX512ER__) || !defined(__AVX512PF__)
|
| 10 |
+
#error "HOST/ARCH doesn't support Knights Landing AVX512 features"
|
| 11 |
+
#endif
|
| 12 |
+
#endif
|
| 13 |
+
|
| 14 |
+
#include <immintrin.h>
|
| 15 |
+
|
| 16 |
+
int main(int argc, char **argv)
|
| 17 |
+
{
|
| 18 |
+
int base[128]={};
|
| 19 |
+
__m512d ad = _mm512_loadu_pd((const __m512d*)argv[argc-1]);
|
| 20 |
+
/* ER */
|
| 21 |
+
__m512i a = _mm512_castpd_si512(_mm512_exp2a23_pd(ad));
|
| 22 |
+
/* PF */
|
| 23 |
+
_mm512_mask_prefetch_i64scatter_pd(base, _mm512_cmpeq_epi64_mask(a, a), a, 1, _MM_HINT_T1);
|
| 24 |
+
return base[0];
|
| 25 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_avx512_knm.c
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#if defined(DETECT_FEATURES) && defined(__INTEL_COMPILER)
|
| 2 |
+
/*
|
| 3 |
+
* Unlike GCC and CLANG, Intel Compiler exposes all supported intrinsics,
|
| 4 |
+
* whether or not the build options for those features are specified.
|
| 5 |
+
* Therefore, we must test #definitions of CPU features when option native/host
|
| 6 |
+
* is enabled via `--cpu-baseline` or through env var `CFLAGS` otherwise
|
| 7 |
+
* the test will be broken and leads to enable all possible features.
|
| 8 |
+
*/
|
| 9 |
+
#if !defined(__AVX5124FMAPS__) || !defined(__AVX5124VNNIW__) || !defined(__AVX512VPOPCNTDQ__)
|
| 10 |
+
#error "HOST/ARCH doesn't support Knights Mill AVX512 features"
|
| 11 |
+
#endif
|
| 12 |
+
#endif
|
| 13 |
+
|
| 14 |
+
#include <immintrin.h>
|
| 15 |
+
|
| 16 |
+
int main(int argc, char **argv)
|
| 17 |
+
{
|
| 18 |
+
__m512i a = _mm512_loadu_si512((const __m512i*)argv[argc-1]);
|
| 19 |
+
__m512 b = _mm512_loadu_ps((const __m512*)argv[argc-2]);
|
| 20 |
+
|
| 21 |
+
/* 4FMAPS */
|
| 22 |
+
b = _mm512_4fmadd_ps(b, b, b, b, b, NULL);
|
| 23 |
+
/* 4VNNIW */
|
| 24 |
+
a = _mm512_4dpwssd_epi32(a, a, a, a, a, NULL);
|
| 25 |
+
/* VPOPCNTDQ */
|
| 26 |
+
a = _mm512_popcnt_epi64(a);
|
| 27 |
+
|
| 28 |
+
a = _mm512_add_epi32(a, _mm512_castps_si512(b));
|
| 29 |
+
return _mm_cvtsi128_si32(_mm512_castsi512_si128(a));
|
| 30 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_avx512_skx.c
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#if defined(DETECT_FEATURES) && defined(__INTEL_COMPILER)
|
| 2 |
+
/*
|
| 3 |
+
* Unlike GCC and CLANG, Intel Compiler exposes all supported intrinsics,
|
| 4 |
+
* whether or not the build options for those features are specified.
|
| 5 |
+
* Therefore, we must test #definitions of CPU features when option native/host
|
| 6 |
+
* is enabled via `--cpu-baseline` or through env var `CFLAGS` otherwise
|
| 7 |
+
* the test will be broken and leads to enable all possible features.
|
| 8 |
+
*/
|
| 9 |
+
#if !defined(__AVX512VL__) || !defined(__AVX512BW__) || !defined(__AVX512DQ__)
|
| 10 |
+
#error "HOST/ARCH doesn't support SkyLake AVX512 features"
|
| 11 |
+
#endif
|
| 12 |
+
#endif
|
| 13 |
+
|
| 14 |
+
#include <immintrin.h>
|
| 15 |
+
|
| 16 |
+
int main(int argc, char **argv)
|
| 17 |
+
{
|
| 18 |
+
__m512i aa = _mm512_abs_epi32(_mm512_loadu_si512((const __m512i*)argv[argc-1]));
|
| 19 |
+
/* VL */
|
| 20 |
+
__m256i a = _mm256_abs_epi64(_mm512_extracti64x4_epi64(aa, 1));
|
| 21 |
+
/* DQ */
|
| 22 |
+
__m512i b = _mm512_broadcast_i32x8(a);
|
| 23 |
+
/* BW */
|
| 24 |
+
b = _mm512_abs_epi16(b);
|
| 25 |
+
return _mm_cvtsi128_si32(_mm512_castsi512_si128(b));
|
| 26 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_avx512_spr.c
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#if defined(DETECT_FEATURES) && defined(__INTEL_COMPILER)
|
| 2 |
+
/*
|
| 3 |
+
* Unlike GCC and CLANG, Intel Compiler exposes all supported intrinsics,
|
| 4 |
+
* whether or not the build options for those features are specified.
|
| 5 |
+
* Therefore, we must test #definitions of CPU features when option native/host
|
| 6 |
+
* is enabled via `--cpu-baseline` or through env var `CFLAGS` otherwise
|
| 7 |
+
* the test will be broken and leads to enable all possible features.
|
| 8 |
+
*/
|
| 9 |
+
#if !defined(__AVX512FP16__)
|
| 10 |
+
#error "HOST/ARCH doesn't support Sapphire Rapids AVX512FP16 features"
|
| 11 |
+
#endif
|
| 12 |
+
#endif
|
| 13 |
+
|
| 14 |
+
#include <immintrin.h>
|
| 15 |
+
|
| 16 |
+
int main(int argc, char **argv)
|
| 17 |
+
{
|
| 18 |
+
/* clang has a bug regarding our spr coode, see gh-23730. */
|
| 19 |
+
#if __clang__
|
| 20 |
+
#error
|
| 21 |
+
#endif
|
| 22 |
+
__m512h a = _mm512_loadu_ph((void*)argv[argc-1]);
|
| 23 |
+
__m512h temp = _mm512_fmadd_ph(a, a, a);
|
| 24 |
+
_mm512_storeu_ph((void*)(argv[argc-1]), temp);
|
| 25 |
+
return 0;
|
| 26 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_avx512cd.c
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#if defined(DETECT_FEATURES) && defined(__INTEL_COMPILER)
|
| 2 |
+
/*
|
| 3 |
+
* Unlike GCC and CLANG, Intel Compiler exposes all supported intrinsics,
|
| 4 |
+
* whether or not the build options for those features are specified.
|
| 5 |
+
* Therefore, we must test #definitions of CPU features when option native/host
|
| 6 |
+
* is enabled via `--cpu-baseline` or through env var `CFLAGS` otherwise
|
| 7 |
+
* the test will be broken and leads to enable all possible features.
|
| 8 |
+
*/
|
| 9 |
+
#ifndef __AVX512CD__
|
| 10 |
+
#error "HOST/ARCH doesn't support AVX512CD"
|
| 11 |
+
#endif
|
| 12 |
+
#endif
|
| 13 |
+
|
| 14 |
+
#include <immintrin.h>
|
| 15 |
+
|
| 16 |
+
int main(int argc, char **argv)
|
| 17 |
+
{
|
| 18 |
+
__m512i a = _mm512_lzcnt_epi32(_mm512_loadu_si512((const __m512i*)argv[argc-1]));
|
| 19 |
+
return _mm_cvtsi128_si32(_mm512_castsi512_si128(a));
|
| 20 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_avx512f.c
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#if defined(DETECT_FEATURES) && defined(__INTEL_COMPILER)
|
| 2 |
+
/*
|
| 3 |
+
* Unlike GCC and CLANG, Intel Compiler exposes all supported intrinsics,
|
| 4 |
+
* whether or not the build options for those features are specified.
|
| 5 |
+
* Therefore, we must test #definitions of CPU features when option native/host
|
| 6 |
+
* is enabled via `--cpu-baseline` or through env var `CFLAGS` otherwise
|
| 7 |
+
* the test will be broken and leads to enable all possible features.
|
| 8 |
+
*/
|
| 9 |
+
#ifndef __AVX512F__
|
| 10 |
+
#error "HOST/ARCH doesn't support AVX512F"
|
| 11 |
+
#endif
|
| 12 |
+
#endif
|
| 13 |
+
|
| 14 |
+
#include <immintrin.h>
|
| 15 |
+
|
| 16 |
+
int main(int argc, char **argv)
|
| 17 |
+
{
|
| 18 |
+
__m512i a = _mm512_abs_epi32(_mm512_loadu_si512((const __m512i*)argv[argc-1]));
|
| 19 |
+
return _mm_cvtsi128_si32(_mm512_castsi512_si128(a));
|
| 20 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_fma4.c
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#include <immintrin.h>
|
| 2 |
+
#ifdef _MSC_VER
|
| 3 |
+
#include <ammintrin.h>
|
| 4 |
+
#else
|
| 5 |
+
#include <x86intrin.h>
|
| 6 |
+
#endif
|
| 7 |
+
|
| 8 |
+
int main(int argc, char **argv)
|
| 9 |
+
{
|
| 10 |
+
__m256 a = _mm256_loadu_ps((const float*)argv[argc-1]);
|
| 11 |
+
a = _mm256_macc_ps(a, a, a);
|
| 12 |
+
return (int)_mm_cvtss_f32(_mm256_castps256_ps128(a));
|
| 13 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_neon_fp16.c
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#ifdef _MSC_VER
|
| 2 |
+
#include <Intrin.h>
|
| 3 |
+
#endif
|
| 4 |
+
#include <arm_neon.h>
|
| 5 |
+
|
| 6 |
+
int main(int argc, char **argv)
|
| 7 |
+
{
|
| 8 |
+
short *src = (short*)argv[argc-1];
|
| 9 |
+
float32x4_t v_z4 = vcvt_f32_f16((float16x4_t)vld1_s16(src));
|
| 10 |
+
return (int)vgetq_lane_f32(v_z4, 0);
|
| 11 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_neon_vfpv4.c
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#ifdef _MSC_VER
|
| 2 |
+
#include <Intrin.h>
|
| 3 |
+
#endif
|
| 4 |
+
#include <arm_neon.h>
|
| 5 |
+
|
| 6 |
+
int main(int argc, char **argv)
|
| 7 |
+
{
|
| 8 |
+
float *src = (float*)argv[argc-1];
|
| 9 |
+
float32x4_t v1 = vdupq_n_f32(src[0]);
|
| 10 |
+
float32x4_t v2 = vdupq_n_f32(src[1]);
|
| 11 |
+
float32x4_t v3 = vdupq_n_f32(src[2]);
|
| 12 |
+
int ret = (int)vgetq_lane_f32(vfmaq_f32(v1, v2, v3), 0);
|
| 13 |
+
#ifdef __aarch64__
|
| 14 |
+
double *src2 = (double*)argv[argc-2];
|
| 15 |
+
float64x2_t vd1 = vdupq_n_f64(src2[0]);
|
| 16 |
+
float64x2_t vd2 = vdupq_n_f64(src2[1]);
|
| 17 |
+
float64x2_t vd3 = vdupq_n_f64(src2[2]);
|
| 18 |
+
ret += (int)vgetq_lane_f64(vfmaq_f64(vd1, vd2, vd3), 0);
|
| 19 |
+
#endif
|
| 20 |
+
return ret;
|
| 21 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_popcnt.c
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#if defined(DETECT_FEATURES) && defined(__INTEL_COMPILER)
|
| 2 |
+
/*
|
| 3 |
+
* Unlike GCC and CLANG, Intel Compiler exposes all supported intrinsics,
|
| 4 |
+
* whether or not the build options for those features are specified.
|
| 5 |
+
* Therefore, we must test #definitions of CPU features when option native/host
|
| 6 |
+
* is enabled via `--cpu-baseline` or through env vr `CFLAGS` otherwise
|
| 7 |
+
* the test will be broken and leads to enable all possible features.
|
| 8 |
+
*/
|
| 9 |
+
#if !defined(__SSE4_2__) && !defined(__POPCNT__)
|
| 10 |
+
#error "HOST/ARCH doesn't support POPCNT"
|
| 11 |
+
#endif
|
| 12 |
+
#endif
|
| 13 |
+
|
| 14 |
+
#ifdef _MSC_VER
|
| 15 |
+
#include <nmmintrin.h>
|
| 16 |
+
#else
|
| 17 |
+
#include <popcntintrin.h>
|
| 18 |
+
#endif
|
| 19 |
+
|
| 20 |
+
int main(int argc, char **argv)
|
| 21 |
+
{
|
| 22 |
+
// To make sure popcnt instructions are generated
|
| 23 |
+
// and been tested against the assembler
|
| 24 |
+
unsigned long long a = *((unsigned long long*)argv[argc-1]);
|
| 25 |
+
unsigned int b = *((unsigned int*)argv[argc-2]);
|
| 26 |
+
|
| 27 |
+
#if defined(_M_X64) || defined(__x86_64__)
|
| 28 |
+
a = _mm_popcnt_u64(a);
|
| 29 |
+
#endif
|
| 30 |
+
b = _mm_popcnt_u32(b);
|
| 31 |
+
return (int)a + b;
|
| 32 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_sse.c
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#if defined(DETECT_FEATURES) && defined(__INTEL_COMPILER)
|
| 2 |
+
/*
|
| 3 |
+
* Unlike GCC and CLANG, Intel Compiler exposes all supported intrinsics,
|
| 4 |
+
* whether or not the build options for those features are specified.
|
| 5 |
+
* Therefore, we must test #definitions of CPU features when option native/host
|
| 6 |
+
* is enabled via `--cpu-baseline` or through env var `CFLAGS` otherwise
|
| 7 |
+
* the test will be broken and leads to enable all possible features.
|
| 8 |
+
*/
|
| 9 |
+
#ifndef __SSE__
|
| 10 |
+
#error "HOST/ARCH doesn't support SSE"
|
| 11 |
+
#endif
|
| 12 |
+
#endif
|
| 13 |
+
|
| 14 |
+
#include <xmmintrin.h>
|
| 15 |
+
|
| 16 |
+
int main(void)
|
| 17 |
+
{
|
| 18 |
+
__m128 a = _mm_add_ps(_mm_setzero_ps(), _mm_setzero_ps());
|
| 19 |
+
return (int)_mm_cvtss_f32(a);
|
| 20 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_sse2.c
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#if defined(DETECT_FEATURES) && defined(__INTEL_COMPILER)
|
| 2 |
+
/*
|
| 3 |
+
* Unlike GCC and CLANG, Intel Compiler exposes all supported intrinsics,
|
| 4 |
+
* whether or not the build options for those features are specified.
|
| 5 |
+
* Therefore, we must test #definitions of CPU features when option native/host
|
| 6 |
+
* is enabled via `--cpu-baseline` or through env var `CFLAGS` otherwise
|
| 7 |
+
* the test will be broken and leads to enable all possible features.
|
| 8 |
+
*/
|
| 9 |
+
#ifndef __SSE2__
|
| 10 |
+
#error "HOST/ARCH doesn't support SSE2"
|
| 11 |
+
#endif
|
| 12 |
+
#endif
|
| 13 |
+
|
| 14 |
+
#include <emmintrin.h>
|
| 15 |
+
|
| 16 |
+
int main(void)
|
| 17 |
+
{
|
| 18 |
+
__m128i a = _mm_add_epi16(_mm_setzero_si128(), _mm_setzero_si128());
|
| 19 |
+
return _mm_cvtsi128_si32(a);
|
| 20 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_sse3.c
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#if defined(DETECT_FEATURES) && defined(__INTEL_COMPILER)
|
| 2 |
+
/*
|
| 3 |
+
* Unlike GCC and CLANG, Intel Compiler exposes all supported intrinsics,
|
| 4 |
+
* whether or not the build options for those features are specified.
|
| 5 |
+
* Therefore, we must test #definitions of CPU features when option native/host
|
| 6 |
+
* is enabled via `--cpu-baseline` or through env var `CFLAGS` otherwise
|
| 7 |
+
* the test will be broken and leads to enable all possible features.
|
| 8 |
+
*/
|
| 9 |
+
#ifndef __SSE3__
|
| 10 |
+
#error "HOST/ARCH doesn't support SSE3"
|
| 11 |
+
#endif
|
| 12 |
+
#endif
|
| 13 |
+
|
| 14 |
+
#include <pmmintrin.h>
|
| 15 |
+
|
| 16 |
+
int main(void)
|
| 17 |
+
{
|
| 18 |
+
__m128 a = _mm_hadd_ps(_mm_setzero_ps(), _mm_setzero_ps());
|
| 19 |
+
return (int)_mm_cvtss_f32(a);
|
| 20 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_sse41.c
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#if defined(DETECT_FEATURES) && defined(__INTEL_COMPILER)
|
| 2 |
+
/*
|
| 3 |
+
* Unlike GCC and CLANG, Intel Compiler exposes all supported intrinsics,
|
| 4 |
+
* whether or not the build options for those features are specified.
|
| 5 |
+
* Therefore, we must test #definitions of CPU features when option native/host
|
| 6 |
+
* is enabled via `--cpu-baseline` or through env var `CFLAGS` otherwise
|
| 7 |
+
* the test will be broken and leads to enable all possible features.
|
| 8 |
+
*/
|
| 9 |
+
#ifndef __SSE4_1__
|
| 10 |
+
#error "HOST/ARCH doesn't support SSE41"
|
| 11 |
+
#endif
|
| 12 |
+
#endif
|
| 13 |
+
|
| 14 |
+
#include <smmintrin.h>
|
| 15 |
+
|
| 16 |
+
int main(void)
|
| 17 |
+
{
|
| 18 |
+
__m128 a = _mm_floor_ps(_mm_setzero_ps());
|
| 19 |
+
return (int)_mm_cvtss_f32(a);
|
| 20 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_ssse3.c
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#if defined(DETECT_FEATURES) && defined(__INTEL_COMPILER)
|
| 2 |
+
/*
|
| 3 |
+
* Unlike GCC and CLANG, Intel Compiler exposes all supported intrinsics,
|
| 4 |
+
* whether or not the build options for those features are specified.
|
| 5 |
+
* Therefore, we must test #definitions of CPU features when option native/host
|
| 6 |
+
* is enabled via `--cpu-baseline` or through env var `CFLAGS` otherwise
|
| 7 |
+
* the test will be broken and leads to enable all possible features.
|
| 8 |
+
*/
|
| 9 |
+
#ifndef __SSSE3__
|
| 10 |
+
#error "HOST/ARCH doesn't support SSSE3"
|
| 11 |
+
#endif
|
| 12 |
+
#endif
|
| 13 |
+
|
| 14 |
+
#include <tmmintrin.h>
|
| 15 |
+
|
| 16 |
+
int main(void)
|
| 17 |
+
{
|
| 18 |
+
__m128i a = _mm_hadd_epi16(_mm_setzero_si128(), _mm_setzero_si128());
|
| 19 |
+
return (int)_mm_cvtsi128_si32(a);
|
| 20 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_vsx3.c
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#ifndef __VSX__
|
| 2 |
+
#error "VSX is not supported"
|
| 3 |
+
#endif
|
| 4 |
+
#include <altivec.h>
|
| 5 |
+
|
| 6 |
+
typedef __vector unsigned int v_uint32x4;
|
| 7 |
+
|
| 8 |
+
int main(void)
|
| 9 |
+
{
|
| 10 |
+
v_uint32x4 z4 = (v_uint32x4){0, 0, 0, 0};
|
| 11 |
+
z4 = vec_absd(z4, z4);
|
| 12 |
+
return (int)vec_extract(z4, 0);
|
| 13 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_vx.c
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#if (__VEC__ < 10301) || (__ARCH__ < 11)
|
| 2 |
+
#error VX not supported
|
| 3 |
+
#endif
|
| 4 |
+
|
| 5 |
+
#include <vecintrin.h>
|
| 6 |
+
int main(int argc, char **argv)
|
| 7 |
+
{
|
| 8 |
+
__vector double x = vec_abs(vec_xl(argc, (double*)argv));
|
| 9 |
+
__vector double y = vec_load_len((double*)argv, (unsigned int)argc);
|
| 10 |
+
|
| 11 |
+
x = vec_round(vec_ceil(x) + vec_floor(y));
|
| 12 |
+
__vector bool long long m = vec_cmpge(x, y);
|
| 13 |
+
__vector long long i = vec_signed(vec_sel(x, y, m));
|
| 14 |
+
|
| 15 |
+
return (int)vec_extract(i, 0);
|
| 16 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_vxe.c
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#if (__VEC__ < 10302) || (__ARCH__ < 12)
|
| 2 |
+
#error VXE not supported
|
| 3 |
+
#endif
|
| 4 |
+
|
| 5 |
+
#include <vecintrin.h>
|
| 6 |
+
int main(int argc, char **argv)
|
| 7 |
+
{
|
| 8 |
+
__vector float x = vec_nabs(vec_xl(argc, (float*)argv));
|
| 9 |
+
__vector float y = vec_load_len((float*)argv, (unsigned int)argc);
|
| 10 |
+
|
| 11 |
+
x = vec_round(vec_ceil(x) + vec_floor(y));
|
| 12 |
+
__vector bool int m = vec_cmpge(x, y);
|
| 13 |
+
x = vec_sel(x, y, m);
|
| 14 |
+
|
| 15 |
+
// need to test the existence of intrin "vflls" since vec_doublee
|
| 16 |
+
// is vec_doublee maps to wrong intrin "vfll".
|
| 17 |
+
// see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100871
|
| 18 |
+
#if defined(__GNUC__) && !defined(__clang__)
|
| 19 |
+
__vector long long i = vec_signed(__builtin_s390_vflls(x));
|
| 20 |
+
#else
|
| 21 |
+
__vector long long i = vec_signed(vec_doublee(x));
|
| 22 |
+
#endif
|
| 23 |
+
|
| 24 |
+
return (int)vec_extract(i, 0);
|
| 25 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/cpu_xop.c
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#include <immintrin.h>
|
| 2 |
+
#ifdef _MSC_VER
|
| 3 |
+
#include <ammintrin.h>
|
| 4 |
+
#else
|
| 5 |
+
#include <x86intrin.h>
|
| 6 |
+
#endif
|
| 7 |
+
|
| 8 |
+
int main(void)
|
| 9 |
+
{
|
| 10 |
+
__m128i a = _mm_comge_epu32(_mm_setzero_si128(), _mm_setzero_si128());
|
| 11 |
+
return _mm_cvtsi128_si32(a);
|
| 12 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/extra_avx512bw_mask.c
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#include <immintrin.h>
|
| 2 |
+
/**
|
| 3 |
+
* Test BW mask operations due to:
|
| 4 |
+
* - MSVC has supported it since vs2019 see,
|
| 5 |
+
* https://developercommunity.visualstudio.com/content/problem/518298/missing-avx512bw-mask-intrinsics.html
|
| 6 |
+
* - Clang >= v8.0
|
| 7 |
+
* - GCC >= v7.1
|
| 8 |
+
*/
|
| 9 |
+
int main(void)
|
| 10 |
+
{
|
| 11 |
+
__mmask64 m64 = _mm512_cmpeq_epi8_mask(_mm512_set1_epi8((char)1), _mm512_set1_epi8((char)1));
|
| 12 |
+
m64 = _kor_mask64(m64, m64);
|
| 13 |
+
m64 = _kxor_mask64(m64, m64);
|
| 14 |
+
m64 = _cvtu64_mask64(_cvtmask64_u64(m64));
|
| 15 |
+
m64 = _mm512_kunpackd(m64, m64);
|
| 16 |
+
m64 = (__mmask64)_mm512_kunpackw((__mmask32)m64, (__mmask32)m64);
|
| 17 |
+
return (int)_cvtmask64_u64(m64);
|
| 18 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/checks/extra_avx512dq_mask.c
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#include <immintrin.h>
|
| 2 |
+
/**
|
| 3 |
+
* Test DQ mask operations due to:
|
| 4 |
+
* - MSVC has supported it since vs2019 see,
|
| 5 |
+
* https://developercommunity.visualstudio.com/content/problem/518298/missing-avx512bw-mask-intrinsics.html
|
| 6 |
+
* - Clang >= v8.0
|
| 7 |
+
* - GCC >= v7.1
|
| 8 |
+
*/
|
| 9 |
+
int main(void)
|
| 10 |
+
{
|
| 11 |
+
__mmask8 m8 = _mm512_cmpeq_epi64_mask(_mm512_set1_epi64(1), _mm512_set1_epi64(1));
|
| 12 |
+
m8 = _kor_mask8(m8, m8);
|
| 13 |
+
m8 = _kxor_mask8(m8, m8);
|
| 14 |
+
m8 = _cvtu32_mask8(_cvtmask8_u32(m8));
|
| 15 |
+
return (int)_cvtmask8_u32(m8);
|
| 16 |
+
}
|
.venv/lib/python3.11/site-packages/numpy/distutils/command/__init__.py
ADDED
|
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""distutils.command
|
| 2 |
+
|
| 3 |
+
Package containing implementation of all the standard Distutils
|
| 4 |
+
commands.
|
| 5 |
+
|
| 6 |
+
"""
|
| 7 |
+
def test_na_writable_attributes_deletion():
|
| 8 |
+
a = np.NA(2)
|
| 9 |
+
attr = ['payload', 'dtype']
|
| 10 |
+
for s in attr:
|
| 11 |
+
assert_raises(AttributeError, delattr, a, s)
|
| 12 |
+
|
| 13 |
+
|
| 14 |
+
__revision__ = "$Id: __init__.py,v 1.3 2005/05/16 11:08:49 pearu Exp $"
|
| 15 |
+
|
| 16 |
+
distutils_all = [ #'build_py',
|
| 17 |
+
'clean',
|
| 18 |
+
'install_clib',
|
| 19 |
+
'install_scripts',
|
| 20 |
+
'bdist',
|
| 21 |
+
'bdist_dumb',
|
| 22 |
+
'bdist_wininst',
|
| 23 |
+
]
|
| 24 |
+
|
| 25 |
+
__import__('distutils.command', globals(), locals(), distutils_all)
|
| 26 |
+
|
| 27 |
+
__all__ = ['build',
|
| 28 |
+
'config_compiler',
|
| 29 |
+
'config',
|
| 30 |
+
'build_src',
|
| 31 |
+
'build_py',
|
| 32 |
+
'build_ext',
|
| 33 |
+
'build_clib',
|
| 34 |
+
'build_scripts',
|
| 35 |
+
'install',
|
| 36 |
+
'install_data',
|
| 37 |
+
'install_headers',
|
| 38 |
+
'install_lib',
|
| 39 |
+
'bdist_rpm',
|
| 40 |
+
'sdist',
|
| 41 |
+
] + distutils_all
|
.venv/lib/python3.11/site-packages/numpy/distutils/command/__pycache__/__init__.cpython-311.pyc
ADDED
|
Binary file (1.28 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/command/__pycache__/autodist.cpython-311.pyc
ADDED
|
Binary file (5.27 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/command/__pycache__/bdist_rpm.cpython-311.pyc
ADDED
|
Binary file (1.31 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/command/__pycache__/build.cpython-311.pyc
ADDED
|
Binary file (3.09 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/command/__pycache__/build_clib.cpython-311.pyc
ADDED
|
Binary file (21.3 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/command/__pycache__/build_ext.cpython-311.pyc
ADDED
|
Binary file (32.2 kB). View file
|
|
|
.venv/lib/python3.11/site-packages/numpy/distutils/command/__pycache__/build_py.cpython-311.pyc
ADDED
|
Binary file (2.16 kB). View file
|
|
|