Skip to content

ENH: Use AVX512-FP16 SVML content for float16 umath functions #23351

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
seiko2plus merged 8 commits into from
Jan 22, 2024
Merged

ENH: Use AVX512-FP16 SVML content for float16 umath functions #23351

Changes from 1 commit
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
55fee61
ENH: Vectorize FP16 math functions on Intel Sapphire Rapids
Jul 25, 2022
c8c0a18
Fix linter errors
Mar 14, 2023
8d2af7f
TST: Adjust for ULP errors in FP16 math functions
Mar 21, 2023
efc6af3
MAINT: Correct svml function signature
Mar 23, 2023
f95896d
Add half precision svml files to meson
Jan 18, 2024
0a0ac43
BUG: Fix function declaration for half precision svml funcs
Jan 19, 2024
34ec3fb
BLD: Conditionally add SVML FP16 objects
Jan 19, 2024
bc27299
BLD: combine variable names into one
Jan 22, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
BUG: Fix function declaration for half precision svml funcs
  • Loading branch information
Raghuveer Devulapalli
Raghuveer Devulapalli committed Jan 19, 2024
commit 0a0ac43e6d260a1365d2cca9534c7f4e1d26c253
42 changes: 21 additions & 21 deletions numpy/_core/src/common/npy_svml.h
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,25 +1,25 @@
#if NPY_SIMD && defined(NPY_HAVE_AVX512_SPR) && defined(NPY_CAN_LINK_SVML)
extern __m512h __svml_exps32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_exp2s32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_logs32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_log2s32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_log10s32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_expm1s32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_log1ps32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_cbrts32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_sins32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_coss32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_tans32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_asins32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_acoss32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_atans32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_atan2s32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_sinhs32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_coshs32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_tanhs32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_asinhs32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_acoshs32(const npy_half*, npy_half*, npy_intp);
extern __m512h __svml_atanhs32(const npy_half*, npy_half*, npy_intp);
extern void __svml_exps32(const npy_half*, npy_half*, npy_intp);
extern void __svml_exp2s32(const npy_half*, npy_half*, npy_intp);
extern void __svml_logs32(const npy_half*, npy_half*, npy_intp);
extern void __svml_log2s32(const npy_half*, npy_half*, npy_intp);
extern void __svml_log10s32(const npy_half*, npy_half*, npy_intp);
extern void __svml_expm1s32(const npy_half*, npy_half*, npy_intp);
extern void __svml_log1ps32(const npy_half*, npy_half*, npy_intp);
extern void __svml_cbrts32(const npy_half*, npy_half*, npy_intp);
extern void __svml_sins32(const npy_half*, npy_half*, npy_intp);
extern void __svml_coss32(const npy_half*, npy_half*, npy_intp);
extern void __svml_tans32(const npy_half*, npy_half*, npy_intp);
extern void __svml_asins32(const npy_half*, npy_half*, npy_intp);
extern void __svml_acoss32(const npy_half*, npy_half*, npy_intp);
extern void __svml_atans32(const npy_half*, npy_half*, npy_intp);
extern void __svml_atan2s32(const npy_half*, npy_half*, npy_intp);
extern void __svml_sinhs32(const npy_half*, npy_half*, npy_intp);
extern void __svml_coshs32(const npy_half*, npy_half*, npy_intp);
extern void __svml_tanhs32(const npy_half*, npy_half*, npy_intp);
extern void __svml_asinhs32(const npy_half*, npy_half*, npy_intp);
extern void __svml_acoshs32(const npy_half*, npy_half*, npy_intp);
extern void __svml_atanhs32(const npy_half*, npy_half*, npy_intp);
#endif

#if NPY_SIMD && defined(NPY_HAVE_AVX512_SKX) && defined(NPY_CAN_LINK_SVML)
Expand Down