Martin Kroeker
4c25910da0
Merge pull request #2896 from martin-frbg/intrin-double
...
Add compiler flag for SSE4 where available
2020-10-15 11:12:35 +02:00
damonyu
ef8e7d0279
Add the support for RISC-V Vector.
...
Change-Id: Iae7800a32f5af3903c330882cdf6f292d885f266
2020-10-15 16:09:02 +08:00
Martin Kroeker
9b9ee92d5f
Merge pull request #2897 from Qiyu8/usimd-double
...
Add double precision universal intrinsics for X86/ARM
2020-10-15 08:38:24 +02:00
Martin Kroeker
ae6ac83991
Revert "add double precision SSE"
2020-10-15 08:37:02 +02:00
Qiyu8
4fac91ef37
adapt arm platform
2020-10-15 11:08:10 +08:00
Qiyu8
bfdf4b56da
Add double precision universal intrinsics for X86/ARM
2020-10-15 10:29:42 +08:00
Martin Kroeker
ebf0470fc2
add sse4.1 for DYNAMIC_ARCH kernels
2020-10-14 20:34:33 +02:00
Martin Kroeker
ca160bb440
Add -msse4.1 when SSE4.1 is supported
2020-10-14 19:18:07 +02:00
Martin Kroeker
c9c3ae07af
Add double precision operations
2020-10-14 18:10:45 +02:00
Martin Kroeker
a897bc3bd2
Merge pull request #99 from xianyi/develop
...
rebase
2020-10-14 18:09:20 +02:00
Martin Kroeker
756802df61
Merge pull request #2890 from martin-frbg/s-d-sum
...
Revert special handling of Windows xNRM2 and enable C+intrinsics kern…
2020-10-14 09:02:03 +02:00
Martin Kroeker
01492decf4
Merge pull request #2895 from martin-frbg/sb-tests
...
Fix remaining build errors related to bfloat16 and cmake
2020-10-14 09:01:16 +02:00
Martin Kroeker
bd0752444a
Merge pull request #2894 from RajalakshmiSR/bf16_packing
...
POWER10: Change the packing format for bfloat16
2020-10-14 08:12:08 +02:00
Martin Kroeker
c1f4f5d4e7
Replace Makefile with simplified version again
2020-10-14 01:08:50 +02:00
Martin Kroeker
75e3a92df6
Add express -mavx and -msse options (and fix a stray = for cooperlake)
2020-10-14 01:01:58 +02:00
Martin Kroeker
2a329baa81
Add the BFLOAT16 functions to cmake builds
2020-10-13 23:21:38 +02:00
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.
2020-10-13 16:05:10 -05:00
Martin Kroeker
4bb73c0171
Rename "HALF" type to "BFLOAT16"
2020-10-13 20:07:19 +02:00
Martin Kroeker
bc5c7f9578
Cleanup
2020-10-13 19:56:09 +02:00
Martin Kroeker
437b7fe261
sh prefix renamed to sb
2020-10-13 19:55:14 +02:00
Martin Kroeker
a0ada4bcb8
Merge pull request #98 from xianyi/develop
...
rebase
2020-10-13 18:50:30 +02:00
Martin Kroeker
602a0c7a69
Merge pull request #2892 from RajalakshmiSR/bf16_make
...
Fix build issues with bfloat16
2020-10-13 18:48:37 +02:00
Rajalakshmi Srinivasaraghavan
b5d30b390d
Fix build issues with bfloat16
...
This patch fixes compilation errors due to recent renaming from SH to SB
with BUILD_BFLOAT16.
2020-10-13 11:00:22 -05:00
Martin Kroeker
137ae618db
Fix typo
2020-10-13 15:02:17 +02:00
Martin Kroeker
9e3cff5cf2
Expressly enable -mavx2 on Zen, SkylakeX and Cooperlake as well
2020-10-13 14:41:25 +02:00
Martin Kroeker
d85b968424
Merge pull request #2891 from martin-frbg/fix-2886
...
Fix several bugs and omissions from the BFLOAT16 rename
2020-10-13 13:46:17 +02:00
Martin Kroeker
5f60a32cac
Add -mssse3 if supported by the hardware
2020-10-13 11:57:04 +02:00
Martin Kroeker
fecedc9c69
Add -mssse3
2020-10-13 11:55:41 +02:00
Martin Kroeker
0eacbca85f
Add Haswell and Zen to temporary sse3 whitelist
2020-10-13 11:42:39 +02:00
Martin Kroeker
6999086a2b
whitelist SANDYBRIDGE for SSE3
2020-10-13 10:32:19 +02:00
Martin Kroeker
9dca578c79
Cleanup
2020-10-13 10:14:08 +02:00
Martin Kroeker
1e7eb7b7a9
Fix typos in currently unused sections
2020-10-13 09:17:15 +02:00
Martin Kroeker
84949754a0
Fix bfloat16 conditional
2020-10-13 09:11:36 +02:00
Martin Kroeker
2ae8785603
Add a POWER9 build with BFLOAT16 enabled
2020-10-13 09:07:50 +02:00
Martin Kroeker
e05af6575e
Fix some overlooked "SHBLAS" entries
2020-10-13 09:05:04 +02:00
Martin Kroeker
c1643006ae
Merge pull request #97 from xianyi/develop
...
rebase
2020-10-13 09:01:49 +02:00
Martin Kroeker
8d2df7d066
Revert special handling of Windows xNRM2 and enable C+intrinsics kernel for SSUM/DSUM
2020-10-13 00:14:29 +02:00
Martin Kroeker
08929430cd
Merge pull request #2886 from martin-frbg/issue_2767
...
Rename "HALF" precision functions (sh prefix) to "BFLOAT16" with "sb" prefix
2020-10-13 00:04:35 +02:00
Martin Kroeker
0c84ffe05f
Merge pull request #2881 from mattip/fninit
...
add fninit to reset fpu registers before assembler routines
2020-10-12 23:50:41 +02:00
Martin Kroeker
cb4274e3ad
Merge pull request #2888 from Qiyu8/usimd-sum
...
Optimize the performance of sum by using universal intrinsics
2020-10-12 23:22:08 +02:00
Matti Picus
403eb513a0
use emms instead, add WIN guards
2020-10-12 18:15:01 +03:00
Martin Kroeker
cb839575ed
Convert the prototypes of the unimplemented BFLOAT16 functions to the new naming scheme
2020-10-12 14:44:33 +02:00
Qiyu8
0ed1f07660
Optimize the performance of sum by using universal intrinsics
2020-10-12 19:48:53 +08:00
Martin Kroeker
bb74dd29db
Restore -msse3
2020-10-12 00:42:05 +02:00
Martin Kroeker
629c497b6c
common_sh.h renamed to common_sb.h
2020-10-12 00:27:11 +02:00
Martin Kroeker
2c552f1074
Change "HALF" and "sh" to "BFLOAT16" and "sb"
2020-10-12 00:11:31 +02:00
Martin Kroeker
7ae9e8960e
Change "HALF" and "sh" to "BFLOAT16" and "sb"
2020-10-12 00:08:29 +02:00
Martin Kroeker
e3a29f6b58
Change "HALF" and "sh" to "BFLOAT16" and "sb"
2020-10-12 00:07:37 +02:00
Martin Kroeker
006c7f6671
Change "HALF" and "sh" to "BFLOAT16" and "sb"
2020-10-12 00:06:06 +02:00
Martin Kroeker
85154c2e18
Change "HALF" and "sh" to "BFLOAT16" and "sb"
2020-10-12 00:05:05 +02:00