cuda : fix bounds check for src0 rows in MMVQ kernel (#2231)

* cuda : fix bounds check for src0 rows in MMVQ kernel

* Update ggml-cuda/mmvq.cu

Co-authored-by: Johannes Gäßler <[email protected]>

---------

Co-authored-by: Johannes Gäßler <[email protected]>

ggml-cuda/mmvq.cu

@@ -75,7 +75,7 @@ static __global__ void mul_mat_vec_q(
             tmp[j][i] = warp_reduce_sum(tmp[j][i]);
         }
 
-        if (threadIdx.x < rows_per_cuda_block) {
+        if (threadIdx.x < rows_per_cuda_block && (rows_per_cuda_block == 1 || row0 + threadIdx.x < nrows_dst)) {
             dst[j*nrows_dst + row0 + threadIdx.x] = tmp[j][threadIdx.x];
         }
     }