diff --git a/ggml-cuda.cu b/ggml-cuda.cu
index 8ba0830d0..7227646ee 100644
--- a/ggml-cuda.cu
+++ b/ggml-cuda.cu
@@ -1485,7 +1485,8 @@ static __global__ void dequantize_mul_mat_vec(const void * vx, const dfloat * y,
 
 template <dequantize_1_kernel_t dequantize_kernel>
 static __global__ void dequantize_mul_mat(
-    const void * vx, const float * y, float * dst, const int ncols_x, const int nrows_x, const int ncols_y, const int nrows_dst) {
+    const void * __restrict__ vx, const float * __restrict__ y, float * __restrict__ dst,
+    const int ncols_x, const int nrows_x, const int ncols_y, const int nrows_dst) {
 
     const int nrows_y = ncols_x;
     const int ncols_dst = ncols_y;