vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

musa: enable building fat binaries, enable unified memory, and disable Flash Attention on QY1 (MTT S80) (#9526)

Browse source 
* mtgpu: add mp_21 support

Signed-off-by: Xiaodong Ye <xiaodong.ye@mthreads.com>

* mtgpu: disable flash attention on qy1 (MTT S80); disable q3_k and mul_mat_batched_cublas

Signed-off-by: Xiaodong Ye <xiaodong.ye@mthreads.com>

* mtgpu: enable unified memory

Signed-off-by: Xiaodong Ye <xiaodong.ye@mthreads.com>

* mtgpu: map cublasOperation_t to mublasOperation_t (sync code to latest)

Signed-off-by: Xiaodong Ye <xiaodong.ye@mthreads.com>

---------

Signed-off-by: Xiaodong Ye <xiaodong.ye@mthreads.com>
This commit is contained in:
R0CKSTAR 2024-09-22 22:55:49 +08:00 • committed by GitHub
parent 912c331d3d
commit c35e586ea5
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
6 changed files with 31 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

2
ggml/src/ggml-cuda/vendors/musa.h vendored
View file

@ -26,6 +26,7 @@
#define cublasSetStream mublasSetStream
#define cublasSgemm mublasSgemm
#define cublasStatus_t mublasStatus_t
#define cublasOperation_t mublasOperation_t
#define cublasGetStatusString mublasStatus_to_string
#define cudaDataType_t musaDataType_t
#define cudaDeviceCanAccessPeer musaDeviceCanAccessPeer
@ -56,6 +57,7 @@
#define cudaLaunchHostFunc musaLaunchHostFunc
#define cudaMalloc musaMalloc
#define cudaMallocHost musaMallocHost
#define cudaMallocManaged musaMallocManaged
#define cudaMemcpy musaMemcpy
#define cudaMemcpyAsync musaMemcpyAsync
#define cudaMemcpyPeerAsync musaMemcpyPeerAsync