diff --git a/ggml/src/ggml-cuda/fattn-common.cuh b/ggml/src/ggml-cuda/fattn-common.cuh
index 40c772578..13c5b0a45 100644
--- a/ggml/src/ggml-cuda/fattn-common.cuh
+++ b/ggml/src/ggml-cuda/fattn-common.cuh
@@ -782,7 +782,7 @@ void launch_fattn(
     const ggml_tensor * K = dst->src[1];
     const ggml_tensor * V = dst->src[2];
 
-    const bool V_is_K_view = V->op == GGML_OP_VIEW && V->src[0] == K && V->data == K->data;
+    const bool V_is_K_view = V->view_src && V->view_offs == 0 && (V->view_src == K || V->view_src == K->view_src);
 
     const ggml_tensor * mask  = dst->src[3];
     const ggml_tensor * sinks = dst->src[4];
diff --git a/ggml/src/ggml-cuda/fattn.cu b/ggml/src/ggml-cuda/fattn.cu
index ba2b96bc3..a5e662418 100644
--- a/ggml/src/ggml-cuda/fattn.cu
+++ b/ggml/src/ggml-cuda/fattn.cu
@@ -247,7 +247,7 @@ static best_fattn_kernel ggml_cuda_get_best_fattn_kernel(const int device, const
         }
     }
 
-    const bool V_is_K_view = V->op == GGML_OP_VIEW && V->src[0] == K && V->data == K->data;
+    const bool V_is_K_view = V->view_src && V->view_offs == 0 && (V->view_src == K || V->view_src == K->view_src);
 
     const int cc = ggml_cuda_info().devices[device].cc;