Skip to content

Add casts to the _mm_prefetch instructions in AVX512 CASUM/ZASUM/CSUM/ZSUM #6133

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
martin-frbg merged 4 commits into from
Oct 17, 2026
+8 −8
Merged

Add casts to the _mm_prefetch instructions in AVX512 CASUM/ZASUM/CSUM/ZSUM #6133

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
682597f
Add casts for the prefetch address
martin-frbg Oct 17, 2026
c25a445
Add casts for the prefetch address
martin-frbg Oct 17, 2026
ef2cfd9
Add casts for the prefetch address
martin-frbg Oct 17, 2026
02c37a8
Add casts for the prefetch address
martin-frbg Oct 17, 2026
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
4 changes: 2 additions & 2 deletions kernel/x86_64/casum_microk_skylakex-2.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,15 +29,15 @@ static FLOAT casum_kernel(BLASLONG n, FLOAT *x)

abs_mask1 = (__m128)_mm_set1_epi32(0x7fffffff);

_mm_prefetch(&x1[0], _MM_HINT_T0);
_mm_prefetch((const char*)&x1[0], _MM_HINT_T0);

if (n2 >= 32){
__m128 x00 = _mm_loadu_ps(&x1[ 0]);
__m128 x01 = _mm_loadu_ps(&x1[ 4]);
__m128 x02 = _mm_loadu_ps(&x1[ 8]);
__m128 x03 = _mm_loadu_ps(&x1[12]);

_mm_prefetch(&x1[16], _MM_HINT_T0);
_mm_prefetch((const char*)&x1[16], _MM_HINT_T0);
__m128 x04 = _mm_loadu_ps(&x1[16]);
__m128 x05 = _mm_loadu_ps(&x1[20]);
__m128 x06 = _mm_loadu_ps(&x1[24]);
Expand Down
4 changes: 2 additions & 2 deletions kernel/x86_64/csum_microk_skylakex-2.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,15 +26,15 @@ static FLOAT casum_kernel(BLASLONG n, FLOAT *x)
accum_12 = _mm_setzero_ps();
accum_13 = _mm_setzero_ps();

_mm_prefetch(&x1[0], _MM_HINT_T0);
_mm_prefetch((const char*)&x1[0], _MM_HINT_T0);

if (n2 >= 32){
__m128 x00 = _mm_loadu_ps(&x1[ 0]);
__m128 x01 = _mm_loadu_ps(&x1[ 4]);
__m128 x02 = _mm_loadu_ps(&x1[ 8]);
__m128 x03 = _mm_loadu_ps(&x1[12]);

_mm_prefetch(&x1[16], _MM_HINT_T0);
_mm_prefetch((const char*)&x1[16], _MM_HINT_T0);
__m128 x04 = _mm_loadu_ps(&x1[16]);
__m128 x05 = _mm_loadu_ps(&x1[20]);
__m128 x06 = _mm_loadu_ps(&x1[24]);
Expand Down
4 changes: 2 additions & 2 deletions kernel/x86_64/zasum_microk_skylakex-2.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,14 +30,14 @@ static FLOAT zasum_kernel(BLASLONG n, FLOAT *x)

abs_mask1 = (__m128d)_mm_set1_epi64x(0x7fffffffffffffff);

_mm_prefetch(&x1[0], _MM_HINT_T0);
_mm_prefetch((const char*)&x1[0], _MM_HINT_T0);
if (n2 >= 16){
__m128d x00 = _mm_loadu_pd(&x1[ 0]);
__m128d x01 = _mm_loadu_pd(&x1[ 2]);
__m128d x02 = _mm_loadu_pd(&x1[ 4]);
__m128d x03 = _mm_loadu_pd(&x1[ 6]);

_mm_prefetch(&x1[8], _MM_HINT_T0);
_mm_prefetch((const char*)&x1[8], _MM_HINT_T0);
__m128d x04 = _mm_loadu_pd(&x1[ 8]);
__m128d x05 = _mm_loadu_pd(&x1[10]);
__m128d x06 = _mm_loadu_pd(&x1[12]);
Expand Down
4 changes: 2 additions & 2 deletions kernel/x86_64/zsum_microk_skylakex-2.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,14 +27,14 @@ static FLOAT zsum_kernel(BLASLONG n, FLOAT *x)
accum_12 = _mm_setzero_pd();
accum_13 = _mm_setzero_pd();

_mm_prefetch(&x1[0], _MM_HINT_T0);
_mm_prefetch((const char*)&x1[0], _MM_HINT_T0);
if (n2 >= 16){
__m128d x00 = _mm_loadu_pd(&x1[ 0]);
__m128d x01 = _mm_loadu_pd(&x1[ 2]);
__m128d x02 = _mm_loadu_pd(&x1[ 4]);
__m128d x03 = _mm_loadu_pd(&x1[ 6]);

_mm_prefetch(&x1[8], _MM_HINT_T0);
_mm_prefetch((const char*)&x1[8], _MM_HINT_T0);
__m128d x04 = _mm_loadu_pd(&x1[ 8]);
__m128d x05 = _mm_loadu_pd(&x1[10]);
__m128d x06 = _mm_loadu_pd(&x1[12]);
Expand Down
Loading