diff --git a/ggml/src/ggml-cuda.cu b/ggml/src/ggml-cuda.cu
index 8a844b02..61d61542 100644
--- a/ggml/src/ggml-cuda.cu
+++ b/ggml/src/ggml-cuda.cu
@@ -2310,9 +2310,11 @@ static bool ggml_cuda_compute_forward(ggml_backend_cuda_context & ctx, struct gg
         case GGML_OP_ARGSORT:
             ggml_cuda_op_argsort(ctx, dst);
             break;
+#if !defined(GGML_DISABLE_FLASH_ATTN)
         case GGML_OP_FLASH_ATTN_EXT:
             ggml_cuda_flash_attn_ext(ctx, dst);
             break;
+#endif
         case GGML_OP_CROSS_ENTROPY_LOSS:
             ggml_cuda_cross_entropy_loss(ctx, dst);
             break;