Qiyu8
bfdf4b56da
Add double precision universal intrinsics for X86/ARM
5 years ago
Martin Kroeker
756802df61
Merge pull request #2890 from martin-frbg/s-d-sum
Revert special handling of Windows xNRM2 and enable C+intrinsics kern…
5 years ago
Martin Kroeker
01492decf4
Merge pull request #2895 from martin-frbg/sb-tests
Fix remaining build errors related to bfloat16 and cmake
5 years ago
Martin Kroeker
bd0752444a
Merge pull request #2894 from RajalakshmiSR/bf16_packing
POWER10: Change the packing format for bfloat16
5 years ago
Martin Kroeker
c1f4f5d4e7
Replace Makefile with simplified version again
5 years ago
Martin Kroeker
75e3a92df6
Add express -mavx and -msse options (and fix a stray = for cooperlake)
5 years ago
Martin Kroeker
2a329baa81
Add the BFLOAT16 functions to cmake builds
5 years ago
Rajalakshmi Srinivasaraghavan
0826d68f93
POWER10: Change the packing format for bfloat16
As the new MMA instructions need the inputs in 4x2 order for bfloat16,
changing the format in copy/packing code. This avoids permute instructions
in the gemm kernel inner loop.
5 years ago
Martin Kroeker
4bb73c0171
Rename "HALF" type to "BFLOAT16"
5 years ago
Martin Kroeker
bc5c7f9578
Cleanup
5 years ago
Martin Kroeker
437b7fe261
sh prefix renamed to sb
5 years ago
Martin Kroeker
a0ada4bcb8
Merge pull request #98 from xianyi/develop
rebase
5 years ago
Martin Kroeker
602a0c7a69
Merge pull request #2892 from RajalakshmiSR/bf16_make
Fix build issues with bfloat16
5 years ago
Rajalakshmi Srinivasaraghavan
b5d30b390d
Fix build issues with bfloat16
This patch fixes compilation errors due to recent renaming from SH to SB
with BUILD_BFLOAT16.
5 years ago
Martin Kroeker
137ae618db
Fix typo
5 years ago
Martin Kroeker
9e3cff5cf2
Expressly enable -mavx2 on Zen, SkylakeX and Cooperlake as well
5 years ago
Martin Kroeker
d85b968424
Merge pull request #2891 from martin-frbg/fix-2886
Fix several bugs and omissions from the BFLOAT16 rename
5 years ago
Martin Kroeker
5f60a32cac
Add -mssse3 if supported by the hardware
5 years ago
Martin Kroeker
fecedc9c69
Add -mssse3
5 years ago
Martin Kroeker
0eacbca85f
Add Haswell and Zen to temporary sse3 whitelist
5 years ago
Martin Kroeker
6999086a2b
whitelist SANDYBRIDGE for SSE3
5 years ago
Martin Kroeker
9dca578c79
Cleanup
5 years ago
Martin Kroeker
1e7eb7b7a9
Fix typos in currently unused sections
5 years ago
Martin Kroeker
84949754a0
Fix bfloat16 conditional
5 years ago
Martin Kroeker
2ae8785603
Add a POWER9 build with BFLOAT16 enabled
5 years ago
Martin Kroeker
e05af6575e
Fix some overlooked "SHBLAS" entries
5 years ago
Martin Kroeker
c1643006ae
Merge pull request #97 from xianyi/develop
rebase
5 years ago
Martin Kroeker
8d2df7d066
Revert special handling of Windows xNRM2 and enable C+intrinsics kernel for SSUM/DSUM
5 years ago
Martin Kroeker
08929430cd
Merge pull request #2886 from martin-frbg/issue_2767
Rename "HALF" precision functions (sh prefix) to "BFLOAT16" with "sb" prefix
5 years ago
Martin Kroeker
0c84ffe05f
Merge pull request #2881 from mattip/fninit
add fninit to reset fpu registers before assembler routines
5 years ago
Martin Kroeker
cb4274e3ad
Merge pull request #2888 from Qiyu8/usimd-sum
Optimize the performance of sum by using universal intrinsics
5 years ago
Matti Picus
403eb513a0
use emms instead, add WIN guards
5 years ago
Martin Kroeker
cb839575ed
Convert the prototypes of the unimplemented BFLOAT16 functions to the new naming scheme
5 years ago
Qiyu8
0ed1f07660
Optimize the performance of sum by using universal intrinsics
5 years ago
Martin Kroeker
bb74dd29db
Restore -msse3
5 years ago
Martin Kroeker
629c497b6c
common_sh.h renamed to common_sb.h
5 years ago
Martin Kroeker
2c552f1074
Change "HALF" and "sh" to "BFLOAT16" and "sb"
5 years ago
Martin Kroeker
7ae9e8960e
Change "HALF" and "sh" to "BFLOAT16" and "sb"
5 years ago
Martin Kroeker
e3a29f6b58
Change "HALF" and "sh" to "BFLOAT16" and "sb"
5 years ago
Martin Kroeker
006c7f6671
Change "HALF" and "sh" to "BFLOAT16" and "sb"
5 years ago
Martin Kroeker
85154c2e18
Change "HALF" and "sh" to "BFLOAT16" and "sb"
5 years ago
Martin Kroeker
ae1ab5bfdf
Change "HALF" and "sh" to "BFLOAT16" and "sb"
5 years ago
Martin Kroeker
052f31bc3c
Change "HALF" and "sh" to "BFLOAT16" and "sb"
5 years ago
Martin Kroeker
3aecafad80
Change "HALF" and "sh" to "BFLOAT16" and "sb"
5 years ago
Martin Kroeker
756062afa5
Rename "HALF" and "sh" to "BFLOAT16" and "sb"
5 years ago
Martin Kroeker
2061f7fdff
Rename "HALF" and "sh" to "BFLOAT16" and "sb"
5 years ago
Martin Kroeker
dc8a1afa63
Rename "HALF" and "sh" to "BFLOAT16" and "sb"
5 years ago
Martin Kroeker
32733ded04
Rename "HALF" and "sh" to "BFLOAT16" and "sb"
5 years ago
Martin Kroeker
3bc8e8c334
Rename "HALF" and "sh" to "BFLOAT16"and "sb"
5 years ago
Martin Kroeker
573508f0ee
Rename common_sh.h to common_sb.h
5 years ago