Beinsezii commited on
Commit
d20d1be
·
verified ·
1 Parent(s): 2c93a44

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +5 -0
README.md ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ All quants supported by mmq
2
+
3
+ https://github.com/ggml-org/llama.cpp/blob/e86f3c22211d9b5c3842e2961a022aac9cdbacad/ggml/src/ggml-cuda/mmq.cu#L269-L294
4
+
5
+ For measuring cublas/rocblas vs mmq perf