Home Explore Help
Sign In
cturan
/
llama.cpp
mirror of https://github.com/cturan/llama.cpp
1
0
Fork 0
Files Issues 0 Wiki
Browse Source

Corrected: changed s13 = src1->nb[3] instead of nb[2] (#18724)

Michael Wand 2 weeks ago
parent
ea23c15990
commit
600a366478
1 changed files with 1 additions and 1 deletions
Unified View Show Diff Stats
  1. 1 1
      ggml/src/ggml-cuda/mmq.cu

+ 1 - 1
ggml/src/ggml-cuda/mmq.cu
View File 

@@ -190,7 +190,7 @@ void ggml_cuda_mul_mat_q(
 
     {
 
     {
 
         const int64_t s11 = src1->nb[1] / ts_src1;
 
         const int64_t s11 = src1->nb[1] / ts_src1;
 
         const int64_t s12 = src1->nb[2] / ts_src1;
 
         const int64_t s12 = src1->nb[2] / ts_src1;
 
-        const int64_t s13 = src1->nb[2] / ts_src1;
 
+        const int64_t s13 = src1->nb[3] / ts_src1;
 
 
 
         if (use_native_mxfp4) {
 
         if (use_native_mxfp4) {
 
             quantize_mmq_mxfp4_cuda(src1_d, ids_src1.get(), src1_q8_1.get(), src0->type, ne10, s11, s12, s13,
 
             quantize_mmq_mxfp4_cuda(src1_d, ids_src1.get(), src1_q8_1.get(), src0->type, ne10, s11, s12, s13,