CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-27 08:21:30 +00:00
Code Issues Packages Projects Releases Wiki Activity

CUDA: Optimize PAD_REFLECT_1D (#15957)

Browse Source 
* CUDA: Optimize PAD_REFLECT_1D
feat: add more test cases for PAD_REFLECT_1D

* use fast_div to improve performance

* Apply suggestion from JohannesGaessler

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>

* Apply suggestion from JohannesGaessler

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>

* optimize

* use a concise expression to further speedup the cuda kernel

---------

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
This commit is contained in:
Bowen Han
2025-09-18 11:26:03 -07:00
committed by GitHub
parent 368560a1e3
commit 38dbdf4c05
3 changed files with 76 additions and 54 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
tests/test-backend-ops.cpp
View File

@@ -6507,6 +6507,7 @@ static std::vector<std::unique_ptr<test_case>> make_test_cases_eval() {
test_cases.emplace_back(new test_pad());
test_cases.emplace_back(new test_pad_ext());
test_cases.emplace_back(new test_pad_reflect_1d());
test_cases.emplace_back(new test_pad_reflect_1d(GGML_TYPE_F32, {3000, 384, 4, 1}));
test_cases.emplace_back(new test_roll());
test_cases.emplace_back(new test_arange());
test_cases.emplace_back(new test_timestep_embedding());
@@ -6645,6 +6646,12 @@ static std::vector<std::unique_ptr<test_case>> make_test_cases_perf() {
test_cases.emplace_back(new test_argmax(GGML_TYPE_F32, {1024, 10, 1, 1}));
test_cases.emplace_back(new test_argmax(GGML_TYPE_F32, {32000, 512, 1, 1}));
test_cases.emplace_back(new test_pad_reflect_1d(GGML_TYPE_F32, {512, 34, 2, 1}));
test_cases.emplace_back(new test_pad_reflect_1d(GGML_TYPE_F32, {3000, 80, 1, 1}));
test_cases.emplace_back(new test_pad_reflect_1d(GGML_TYPE_F32, {3000, 80, 4, 1}));
test_cases.emplace_back(new test_pad_reflect_1d(GGML_TYPE_F32, {3000, 384, 1, 1}));
test_cases.emplace_back(new test_pad_reflect_1d(GGML_TYPE_F32, {3000, 384, 4, 1}));
test_cases.emplace_back(new test_mul_mat(GGML_TYPE_F16, GGML_TYPE_F32, 16416, 1, 128, {8, 1}, {4, 1}, {0, 2, 1, 3}));
test_cases.emplace_back(new test_mul_mat(GGML_TYPE_F16, GGML_TYPE_F32, 128, 1, 16416, {8, 1}, {4, 1}, {0, 1, 2, 3}, true));