1 files changed, 5 insertions, 0 deletions

diff --git a/llama.cpp/ggml/src/ggml-cuda/diag.cuh b/llama.cpp/ggml/src/ggml-cuda/diag.cuh
new file mode 100644
index 0000000..7d73e6a
--- /dev/null
+++ b/llama.cpp/ggml/src/ggml-cuda/diag.cuh
@@ -0,0 +1,5 @@
+#include "common.cuh"
+
+#define CUDA_DIAG_BLOCK_SIZE 256
+
+void ggml_cuda_op_diag(ggml_backend_cuda_context & ctx, ggml_tensor * dst);