CUDA: fix shared memory access condition for mmv (ggerganov#10740)

3Simplex · Dec 9, 2024 · 26a8406 · 26a8406
1 parent c37fb4c
commit 26a8406
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/ggml/src/ggml-cuda/mmv.cu b/ggml/src/ggml-cuda/mmv.cu
@@ -57,7 +57,7 @@ static __global__ void mul_mat_vec(
     if (block_size > WARP_SIZE) {
         buf_iw[tid/WARP_SIZE] = sumf;
         __syncthreads();
-        if (tid > WARP_SIZE) {
+        if (tid >= WARP_SIZE) {
             return;
         }
         sumf = buf_iw[tid];