Skip to content

Optimize ZROT_RVV for the unit-stride case (inc_x = inc_y = 1) #5538

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
merged 1 commit into from
Nov 19, 2025
+9 −23
Merged

Optimize ZROT_RVV for the unit-stride case (inc_x = inc_y = 1) #5538

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
32 changes: 9 additions & 23 deletions kernel/riscv64/zrot_rvv.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -97,33 +97,19 @@ int CNAME(BLASLONG n, FLOAT *x, BLASLONG inc_x, FLOAT *y, BLASLONG inc_y, FLOAT
}
else if(inc_x == 1 && inc_y == 1) {

for (size_t vl; n > 0; n -= vl, x += vl*2, y += vl*2) {
n *= 2;
for (size_t vl; n > 0; n -= vl, x += vl, y += vl) {
vl = VSETVL(n);

vxx2 = VLSEG_FLOAT(x, vl);
vyx2 = VLSEG_FLOAT(y, vl);

vx0 = VGET_VX2(vxx2, 0);
vx1 = VGET_VX2(vxx2, 1);
vy0 = VGET_VX2(vyx2, 0);
vy1 = VGET_VX2(vyx2, 1);

vx0 = VLEV_FLOAT(x, vl);
vy0 = VLEV_FLOAT(y, vl);
vt0 = VFMULVF_FLOAT(vx0, c, vl);
vt0 = VFMACCVF_FLOAT(vt0, s, vy0, vl);
vt1 = VFMULVF_FLOAT(vx1, c, vl);
vt1 = VFMACCVF_FLOAT(vt1, s, vy1, vl);
vy0 = VFMULVF_FLOAT(vy0, c, vl);
vy0 = VFNMSACVF_FLOAT(vy0, s, vx0, vl);
vy1 = VFMULVF_FLOAT(vy1, c, vl);
vy1 = VFNMSACVF_FLOAT(vy1, s, vx1, vl);
vx1 = VFMACCVF_FLOAT(vt0, s, vy0, vl);

vtx2 = VSET_VX2(vtx2, 0, vt0);
vtx2 = VSET_VX2(vtx2, 1, vt1);
vyx2 = VSET_VX2(vyx2, 0, vy0);
vyx2 = VSET_VX2(vyx2, 1, vy1);

VSSEG_FLOAT(x, vtx2, vl);
VSSEG_FLOAT(y, vyx2, vl);
vt1 = VFMULVF_FLOAT(vy0, c, vl);
vy1 = VFNMSACVF_FLOAT(vt1, s, vx0, vl);
VSEV_FLOAT(x, vx1, vl);
VSEV_FLOAT(y, vy1, vl);
}

} else if (inc_x == 1){
Expand Down
Loading