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

CUDA: add FP32 FlashAttention vector kernel (#7188)

Browse source 
* CUDA: add FP32 FlashAttention vector kernel

* fixup! CUDA: add FP32 FlashAttention vector kernel

* fixup! fixup! CUDA: add FP32 FlashAttention vector kernel

* fixup! fixup! fixup! CUDA: add FP32 FlashAttention vector kernel
This commit is contained in:
Johannes Gäßler 2024-05-12 19:40:45 +02:00 • committed by GitHub
parent 6f1b63606f
commit dc685be466
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
9 changed files with 899 additions and 458 deletions
Download patch file Download diff file Expand all files Collapse all files

5
tests/test-backend-ops.cpp
View file

@ -2,6 +2,7 @@
#include <ggml-alloc.h>
#include <ggml-backend.h>
#include <ggml-backend-impl.h>
#include <algorithm>
#include <array>
#include <cfloat>
@ -2173,11 +2174,7 @@ static bool test_backend(ggml_backend_t backend, test_mode mode, const char * op
test_cases.emplace_back(new test_timestep_embedding());
test_cases.emplace_back(new test_leaky_relu());
#if defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)
for (int hs : { 64, 128, }) { // other head sizes not implemented
#else
for (int hs : { 64, 80, 128, 256, }) {
#endif // defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)
for (float max_bias : {0.0f, 8.0f}) {
for (int nh : { 32, }) {
for (int kv : { 512, 1024, }) {