phynexis-deps/OpenBLAS
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merge pull request #5618 from vtjnash/jn/zdot_thunderx2t99-ICE

Browse Source 
arm64: fix clang ICE on Windows for thunderx2t kernels
This commit is contained in:
Martin Kroeker
2026-01-20 14:02:30 +01:00
committed by GitHub
parent a10f535803 a18a4ee08a
commit 413e609f9c
4 changed files with 11 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
kernel/arm64/dznrm2_thunderx2t99_fast.c
+3
View File
@@ -155,7 +155,10 @@ static double nrm2_compute(BLASLONG n, FLOAT *x, BLASLONG inc_x)
" cmp "J", xzr \n"
" beq .Lnrm2_kernel_F1 \n"
/* https://github.com/llvm/llvm-project/issues/149547 */
#if !(defined(__clang__) && defined(OS_WINDOWS))
" .align 5 \n"
#endif
".Lnrm2_kernel_F: \n"
" "KERNEL_F" \n"
" subs "J", "J", #1 \n"
kernel/arm64/scnrm2_thunderx2t99.c
+3
View File
@@ -238,7 +238,10 @@ static double nrm2_compute(BLASLONG n, FLOAT *x, BLASLONG inc_x)
" cmp "J", xzr \n"
" beq 5f //nrm2_kernel_S_BEGIN \n"
/* https://github.com/llvm/llvm-project/issues/149547 */
#if !(defined(__clang__) && defined(OS_WINDOWS))
" .align 5 \n"
#endif
"2: //nrm2_kernel_F: \n"
" "KERNEL_F" \n"
" subs "J", "J", #1 \n"
kernel/arm64/zdot_thunderx2t99.c
+2 -1
View File
@@ -243,7 +243,8 @@ static void zdot_compute(BLASLONG n, FLOAT *x, BLASLONG inc_x, FLOAT *y, BLASLON
" asr "J", "N", #"N_DIV_SHIFT" \n"
" cmp "J", xzr \n"
" beq 3f //dot_kernel_F1 \n"
#ifndef _MSC_VER
/* https://github.com/llvm/llvm-project/issues/149547 */
#if !(defined(__clang__) && defined(OS_WINDOWS))
" .align 5 \n"
#endif
"2: //dot_kernel_F: \n"
kernel/arm64/zsum_thunderx2t99.c
+3
View File
@@ -136,7 +136,10 @@ static FLOAT zasum_compute(BLASLONG n, FLOAT *x, BLASLONG inc_x)
" cmp "J", xzr \n"
" beq 3f //asum_kernel_F1 \n"
/* https://github.com/llvm/llvm-project/issues/149547 */
#if !(defined(__clang__) && defined(OS_WINDOWS))
".align 5 \n"
#endif
"2: //asum_kernel_F16: \n"
" "KERNEL_F16" \n"
" subs "J", "J", #1 \n"