root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-25 02:44:36 +00:00
Code Issues Actions 9 Packages Projects Releases Wiki Activity

ggml : avoid conv 2d kernel round up

Browse Source
This commit is contained in:
Georgi Gerganov 2023-06-26 21:03:59 +03:00
parent b853d45601
commit c824d2e368
No known key found for this signature in database
GPG Key ID: 449E073F9DC10735
1 changed files with 1 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
ggml.c
View File

@ -13508,8 +13508,7 @@ static void ggml_compute_forward_conv_2d_sk_p0_f16_f32(
const int nk1 = ne01; const int nk1 = ne01;
// size of the convolution row - the kernel size unrolled across all channels // size of the convolution row - the kernel size unrolled across all channels
// round-up so it is more suitable for SIMD const int ew0 = nk0*nk1*ne02;
const int ew0 = ggml_up32(nk0*nk1*ne02);
GGML_ASSERT(nb00 == sizeof(ggml_fp16_t)); GGML_ASSERT(nb00 == sizeof(ggml_fp16_t));
GGML_ASSERT(nb10 == sizeof(float)); GGML_ASSERT(nb10 == sizeof(float));