root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-11-11 13:30:35 +00:00
Code Issues Actions 11 Packages Projects Releases Wiki Activity

Fix backend test for ssm_conv CUDA op not working

Browse Source
This commit is contained in:
Jan Ploski 2024-06-02 20:32:33 +02:00
parent 64fbd320ef
commit 12c913c52c
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
ggml/src/ggml-cuda/ssm_conv.cu
View File

@ -32,7 +32,7 @@ static __global__ void ssm_conv_f32(
float * s = (float *) wdata_f32 + nc*dr*ith;
for (int i3 = 0; i3 < n_s; ++i3) {
float * s0 = (float *) ((char *) src0 + ir0*src0_nb1) + i3*src0_nb2; // {d_conv, d_inner, n_s}
float * s0 = (float *) ((char *) src0 + ir0*src0_nb1 + i3*src0_nb2); // {d_conv, d_inner, n_s}
// copy the state into working memory
// can't use memcpy because (d_conv) != (d_conv - 1)