mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-10-24 22:00:35 +02:00
Johannes GäßlerandDiego Devesa
864a0b67a6
* CUDA: use mma PTX instructions for FlashAttention * __shfl_sync workaround for movmatrix * add __shfl_sync to HIP Co-authored-by: Diego Devesa <slarengh@gmail.com>
4 lines
115 B
Plaintext
4 lines
115 B
Plaintext
#include "common.cuh"
|
|
|
|
void ggml_cuda_flash_attn_ext_wmma_f16(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|