Martin Kroeker
4d08156266
Use the generic C kernel for DNRM2
2026-01-11 21:58:31 +01:00
Martin Kroeker
d1de282a4e
Improve the precision of S/CNRM2 by summing in double precision
2026-01-11 13:04:00 +01:00
Martin Kroeker
c040d5ed86
Merge pull request #5591 from quic/topic/ssyr2k_direct_sme1
...
Support for SME1 based ssyr2k_direct kernel for cblas_ssyr2k level 3 API
2026-01-08 15:47:38 +01:00
Zhiqing xie
6939a43c3b
Support for SME1 based ssyr2k_direct kernel for cblas_ssyr2k level 3 API
2026-01-08 11:09:04 +08:00
Amrita H S
b53d18b3ad
Fixing warning messages in dgemm and dgemv kernels
...
Signed-off-by: Amrita H S <amritahs@linux.vnet.ibm.com >
2026-01-06 10:20:56 -06:00
Rajalakshmi Srinivasaraghavan
2283fcbbe7
POWER10: Reduce sgemm loop unrolling
...
With GCC 14, unnecessary move and lxvp instructions appear when unrolling the inner loop for larger sizes.
Reducing the loop unroll factor restores performance to GCC 11.
2026-01-04 17:01:01 -06:00
Martin Kroeker
d39b77748f
Make .align conditional on not being on WoA and strip CRLF endings
2025-12-24 20:00:45 +01:00
Martin Kroeker
ac2c66321d
remove special handling of C/ZDOT for LLVM on WoA
2025-12-19 17:04:21 +01:00
Martin Kroeker
cfa28bcf71
Support compilation with LLVM for Windows on Arm
2025-12-19 17:00:47 +01:00
pengxu
f6533ccea0
Fix floating point registers ld/st bug of Loongarch
2025-12-03 10:52:59 +08:00
Martin Kroeker
d6b25c43c6
Merge pull request #5542 from abhishek-iitmadras/abhishek_new_tt_a64fx
...
[A64FX]: add tt for a64fx dot
2025-11-23 22:55:51 +01:00
Martin Kroeker
f7b7296bff
Fix compilation with LLVM
2025-11-22 16:07:34 +01:00
Abhishek Kumar
a14caf464f
add tt for a64fx dot
...
Signed-off-by: Abhishek Kumar <abhishek.r.kumar@fujitsu.com >
2025-11-20 12:14:17 +05:30
Martin Kroeker
28eeef5bbe
Merge pull request #5538 from CheryDan/riscv/rot
...
Optimize ZROT_RVV for the unit-stride case (inc_x = inc_y = 1)
2025-11-19 07:33:08 +01:00
Martin Kroeker
17f2e94260
Merge pull request #5539 from FRosner/arm64-dot-kernel-refactoring
...
Refactoring: ARM64 dot Kernel: don't call num_cpu_avail twice
2025-11-18 23:25:51 +01:00
Frank Rosner
762ed66c72
Refactoring: ARM64 dot Kernel: don't call num_cpu_avail twice
2025-11-18 15:34:51 +01:00
daichengrong
98a8230dee
Optimize ZROT_RVV for the unit-stride case (inc_x = inc_y = 1)
2025-11-18 17:34:27 +08:00
mayeut
39d5e44723
fix: dot_kernel_sve "n" usage & clobber list
2025-11-17 21:53:51 +01:00
Martin Kroeker
f2d010de12
Merge pull request #5512 from quic/topic/ssyrk_direct_sme1
...
Support for SME1 based ssyrk_direct kernel for cblas_ssyrk level 3 API
2025-11-06 14:06:43 -08:00
Chip Kerchner
00a7336fc9
Missing one gemv conversion.
2025-11-04 22:27:53 +00:00
Chip Kerchner
edf2e5900c
Prevent possible conversion from bfloat16 to __bf16.
2025-11-04 21:00:37 +00:00
Martin Kroeker
0c59ae0b45
Merge pull request #5453 from pratiklp00/dgemm_optimization
...
Dgemm loop unroll and 4x1, 4x2 dgemv VSX implementation for power10.
2025-10-28 16:51:41 -07:00
Martin Kroeker
585e6d0680
Merge pull request #5515 from iha-taisei/feature/ger_unroll
...
Improve single-thread performance of [SD]GER on A64FX and Neoverse V1
2025-10-24 08:17:06 -07:00
Iha, Taisei
cb66aca707
Improve single-thread performance of [SD]GER on A64FX and Neoverse V1
2025-10-22 19:56:14 +09:00
Yichao Yu
3d19d3b60a
Make dummy function have the same linkage as the real one
2025-10-20 12:42:39 -04:00
changjua
43d38d336f
Support for SME1 based ssyrk_direct kernel for cblas_ssyrk level 3 API
2025-10-20 11:35:20 +08:00
pratiklp00
6637352260
remmove spacing
2025-10-14 00:06:04 -05:00
Yichao Yu
b94e9b92ad
Fix compilation on ARM
...
Define a dummy function if SME is not supported, following what sgemm does
2025-10-11 20:28:59 -04:00
Martin Kroeker
e40714cabd
Merge pull request #5450 from quic/topic/strmm_direct_sme1
...
Support for SME1 based strmm_direct kernel for cblas_strmm level 3 API
2025-10-11 15:20:19 -07:00
changjua
644ea07ef9
Support for SME1 based strmm_direct kernel for cblas_strmm level 3 API
2025-10-10 10:48:27 +08:00
pratiklp00
e2399be6d2
add macro
2025-10-08 23:24:41 -05:00
Chip Kerchner
03a83778bb
Tie in SHGEMV for RISC-V.
2025-10-08 14:08:29 +00:00
Martin Kroeker
49eca84eaf
Merge pull request #5478 from martin-frbg/issue5477
...
Change all aligned moves in x86_64 MIN/MAX to unaligned
2025-10-08 02:46:00 -07:00
Martin Kroeker
46fc6c0794
fix unspecified array size in clobber list
2025-10-08 08:23:24 +02:00
Martin Kroeker
064751ee65
Merge pull request #5481 from ChipKerchner/vectorSBGEMV
...
Add SBGEMV and SHGEMV routines to RISC-V
2025-10-07 13:31:03 -07:00
Chip Kerchner
f552040c5d
Fix stride issue.
2025-10-07 17:17:18 +00:00
Chris Sidebottom
37fc3bbca0
Add Infrastructure for SHGEMV
...
This adds all the relevant bits and pieces to add a `shgemv` path as
well as a future `hgemm`/`hgemv` path in a similar model to `sb` and `b`
interfaces.
I've also fixed a few bits and pieces around `shgemm` which didn't build
in a few situations.
2025-10-07 15:03:24 +00:00
Chip Kerchner
aecb7f9537
Change signature of SBGEMV.
2025-10-07 13:14:20 +00:00
Chris Sidebottom
958f721e36
Beta fix for generic gemv T
2025-10-07 10:01:12 +00:00
Chris Sidebottom
578e7dae85
Fix bf16->f32 conversion for NEOVERSEV1 and NEOVERSEN2 targets
...
This fixes an issue originally introduced with the BGEMM kernel.
I've updated the tests to run with `beta=1.0` so as to test loading and
updating from C.
Alongside this, the tests now return sensible return values to reduce
the risk of them being ignored.
Also fixed a bug in `generic/gemv_t.c` resulting in weird outputs for
`bgemv`.
2025-10-06 18:05:58 +00:00
Chip Kerchner
809e1cba8f
Better FP16 vectorized GEMV - 20% faster.
2025-10-06 13:19:03 +00:00
Chip Kerchner
e07a9ae418
Merge branch 'develop' into vectorSBGEMV
2025-10-03 17:13:29 +00:00
Chip Kerchner
588f0e87cc
Add SBGEMV and SHGEMV routines to RISC-V.
2025-10-03 17:09:16 +00:00
Martin Kroeker
b48a089d75
Change all aligned moves to unaligned
2025-10-01 23:36:48 +02:00
Martin Kroeker
e939c6c315
Merge pull request #5471 from quic/topic/ssymm_direct_sme1
...
Support for SME1 based ssymm_direct kernel for cblas_ssymm level 3 API
2025-10-01 06:22:36 -07:00
Chip Kerchner
36f9cb85b1
Fix pre-RVV 1.0.
2025-09-30 22:41:31 +00:00
Chip Kerchner
2d82d144e2
Tranverse matrix data in a cache friendly manner for GEMV_N (RISCV).
2025-09-30 21:22:10 +00:00
Martin Kroeker
aaa5c377bc
Merge pull request #5465 from ChipKerchner/addRVVVectorizedFP16Packing
...
Add vectorized packing for FP16 and BF16 for RISC-V. Reactivate vector packing for FP64 transposed
2025-09-30 09:21:15 -07:00
Rajendra Prasad Matcha
19268471cc
Support for SME1 based ssymm_direct kernel for cblas_ssymm level 3 API
2025-09-30 15:05:33 +05:30
Chip Kerchner
67ddda394e
Merge branch 'develop' into addRVVVectorizedFP16Packing
2025-09-29 13:49:57 +00:00
Martin Kroeker