PR #13759 [ROCm] Add GemmFastGelu CK implementation

[ROCm] Add GemmFastGelu CK implementation #13759

PeixuanZuo merged 17 commits into main from peixuanzuo/gemmfastgelu_ck_2

PeixuanZuo requested a review from

zhangyaobit 3 years ago

PeixuanZuo requested a review from

cloudhan 3 years ago

PeixuanZuo force pushed from af24cef8 to b698bc44 3 years ago

zhangyaobit commented on 2022-11-29

cloudhan commented on 2022-12-01

cloudhan commented on 2022-12-13

PeixuanZuo force pushed from 769febb2 to db98e75f 3 years ago

cloudhan commented on 2022-12-14

cloudhan commented on 2022-12-19

PeixuanZuo force pushed from cc38bbcc to 9a43f2cc 3 years ago

change the gemmfastgelu to make it easier to add ckgemmfastgelu

936d6555

update composable kernel version

1cb6e945

add ckgemmfastgelu and tests

8beffd4f

add ckgemmfastglu to gemmfastgelutunableop

3d9d6bde

update python format

ee15b415

move gemmfastgelu from tunable/ to contrib_ops/

d0b0e10a

update gemm_fast_gelu_test.cc

8ea5aafa

update composable_kernel commit

3235924b

move all ck dependences to onnxruntime_provider_rocm

3175e960

update gemm_fast_gelu_test.cc

61b3b7c4

update gemm_fast_gelu_test.py to enable sort

60ed587d

update gemm_fast_gelu_test.cc

4a461fe6

fix error

1486bc09

fix call column_major::Gemm

61949d68

remove unneeded static_cast

801551aa

use registerOp

b906116c

update ck version to 0345963eef4f92e9c5eab608bb8557b5463a1dcb

11e4b871

PeixuanZuo force pushed from 390363cf to 11e4b871 3 years ago

cloudhan approved these changes on 2023-01-05

PeixuanZuo merged 4eac0db3 into main 3 years ago

PeixuanZuo deleted the peixuanzuo/gemmfastgelu_ck_2 branch 3 years ago

Reviewers

cloudhan

zhangyaobit

Assignees

No one assigned

Labels

None yet

Milestone

No milestone

onnxruntime [ROCm] Add GemmFastGelu CK implementation #13759 Merged

[ROCm] Add GemmFastGelu CK implementation #13759

onnxruntime
[ROCm] Add GemmFastGelu CK implementation
#13759

Merged