CUDA: mmq CLI option, fixed mmq build issues (#2453)

1 files changed, 1 insertions, 0 deletions

diff --git a/llama.h b/llama.h
index df46f9b..fa1977f 100644
--- a/llama.h
+++ b/llama.h
@@ -108,6 +108,7 @@ extern "C" {
 
         // Keep the booleans together to avoid misalignment during copy-by-value.
         bool low_vram;   // if true, reduce VRAM usage at the cost of performance
+        bool mul_mat_q;  // if true, use experimental mul_mat_q kernels
         bool f16_kv;     // use fp16 for KV cache
         bool logits_all; // the llama_eval() call computes all logits, not just the last one
         bool vocab_only; // only load the vocabulary, no weights