Skip to content

Implementing SVE in [SD]AXPY Kernels for A64FX and Graviton3E #5426

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.

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

Open
wants to merge 2 commits into
base: develop
Choose a base branch
from
Open
Show file tree
Hide file tree
Changes from all commits
Commits
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
3 changes: 3 additions & 0 deletions kernel/arm64/KERNEL.A64FX
Original file line number Diff line number Diff line change
Expand Up @@ -7,3 +7,6 @@ DGEMVTKERNEL = gemv_t_sve_v4x3.c

DDOTKERNEL = dot_sve_v8.c
SDOTKERNEL = dot_sve_v8.c

SAXPYKERNEL = axpy_sve.c
DAXPYKERNEL = axpy_sve.c
4 changes: 4 additions & 0 deletions kernel/arm64/KERNEL.NEOVERSEV1
Original file line number Diff line number Diff line change
Expand Up @@ -32,6 +32,10 @@ SGEMVNKERNEL = gemv_n_sve_v1x3.c
DGEMVNKERNEL = gemv_n_sve_v1x3.c
SGEMVTKERNEL = gemv_t_sve_v1x3.c
DGEMVTKERNEL = gemv_t_sve_v1x3.c

SAXPYKERNEL = axpy_sve.c
DAXPYKERNEL = axpy_sve.c
Copy link
Contributor

@snadampal snadampal Aug 21, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

since you have used the SVL for the implementation instead of hardcoding the vector width, the kernel should work on NEOVERSEV2 as well. Please check this on Graviton4 and add it to KERNEL.NEOVERSEV2 as well.


ifeq ($(BUILD_BFLOAT16), 1)
BGEMM_BETA = bgemm_beta_neon.c
BGEMMKERNEL = bgemm_kernel_2vlx4_neoversev1.c
Expand Down
86 changes: 86 additions & 0 deletions kernel/arm64/axpy_sve.c
Original file line number Diff line number Diff line change
@@ -0,0 +1,86 @@
/***************************************************************************
Copyright (c) 2025, The OpenBLAS Project
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are
met:
1. Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
2. Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in
the documentation and/or other materials provided with the
distribution.
3. Neither the name of the OpenBLAS project nor the names of
its contributors may be used to endorse or promote products
derived from this software without specific prior written
permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE
USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*****************************************************************************/

#include <arm_sve.h>

#include "common.h"

#ifdef DOUBLE
#define SV_TYPE svfloat64_t
#define SV_COUNT svcntd
#define SV_DUP svdup_f64
#define SV_WHILE svwhilelt_b64_s64
#define SV_TRUE svptrue_b64
#else
#define SV_TYPE svfloat32_t
#define SV_COUNT svcntw
#define SV_DUP svdup_f32
#define SV_WHILE svwhilelt_b32_s64
#define SV_TRUE svptrue_b32
#endif

int CNAME(BLASLONG n, BLASLONG dummy0, BLASLONG dummy1, FLOAT da, FLOAT *x, BLASLONG inc_x, FLOAT *y, BLASLONG inc_y, FLOAT *dummy, BLASLONG dummy2) {
BLASLONG i = 0;
BLASLONG ix = 0, iy = 0;
BLASLONG sve_size = SV_COUNT();

if (n < 0) return (0);
if (da == 0.0) return (0);
Copy link
Contributor

@snadampal snadampal Aug 21, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

why can't these two checks be combined into one?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thank you for your comments.
There was another way you mentioned, but I followed kernel/arm/axpy.c#L45-L46.


if (inc_x == 1 && inc_y == 1) {
SV_TYPE da_vec = SV_DUP(da);
for (i = 0; i + sve_size - 1 < n; i += sve_size) {
SV_TYPE x_vec = svld1(SV_TRUE(), &x[i]);
SV_TYPE y_vec = svld1(SV_TRUE(), &y[i]);
y_vec = svmla_x(SV_TRUE(), y_vec, da_vec, x_vec);
svst1(SV_TRUE(), &y[i], y_vec);
}

if (i < n) {
svbool_t pg = SV_WHILE(i, n);
SV_TYPE x_vec = svld1(pg, &x[i]);
SV_TYPE y_vec = svld1(pg, &y[i]);
y_vec = svmla_x(pg, y_vec, da_vec, x_vec);
svst1(pg, &y[i], y_vec);
}
return (0);
}

while (i < n) {
y[iy] += da * x[ix];
ix += inc_x;
iy += inc_y;
i++;
}

return (0);
}
Loading