CUDA: add roll (#14919)

* CUDA: add roll * Make everything const, use __restrict__
2025-07-29 14:45:18 +08:00
parent 8ad7b3e65b
commit 0a5036bee9
3 changed files with 81 additions and 0 deletions
--- a/ggml/src/ggml-cuda/roll.cuh
+++ b/ggml/src/ggml-cuda/roll.cuh
@@ -0,0 +1,5 @@
+#include "common.cuh"
+
+#define CUDA_ROLL_BLOCK_SIZE 256
+
+void ggml_cuda_op_roll(ggml_backend_cuda_context & ctx, ggml_tensor * dst);