Skip to content

Commit

Permalink
[ hgemm ] Use aligned memory allocation in transpose / padding gemm
Browse files Browse the repository at this point in the history
- Using unaligned memory may invoke SIGSEGV

**Self evaluation:**
1. Build test:     [X]Passed [ ]Failed [ ]Skipped
2. Run test:     [X]Passed [ ]Failed [ ]Skipped

Signed-off-by: skykongkong8 <[email protected]>
  • Loading branch information
skykongkong8 authored and jijoongmoon committed Jul 2, 2024
1 parent dc38b83 commit 2ff7d98
Showing 1 changed file with 0 additions and 1 deletion.
1 change: 0 additions & 1 deletion nntrainer/tensor/hgemm/hgemm.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -1338,7 +1338,6 @@ void hgemm_transB(const __fp16 *A, const __fp16 *B, float *C, unsigned int M,
void hgemm_transA(const __fp16 *A, const __fp16 *B, float *C, unsigned int M,
unsigned int N, unsigned int K, float alpha, float beta) {
__fp16 *A_T = alignedMalloc(M * K);

transpose_neon<__fp16>(K, M, A, M, A_T, K);

hgemm_noTrans(A_T, B, C, M, N, K, alpha, beta);
Expand Down

0 comments on commit 2ff7d98

Please sign in to comment.