onnxruntime
[CUDA] Rollback TileMemcpy and TileBatchedMemcpy when Block Size is Small
#11187

Merged

[CUDA] Rollback TileMemcpy and TileBatchedMemcpy when Block Size is Small #11187

Lafi7e merged 5 commits into master from weicwang/tile_perf

switch kernel impl by block size

61019b5e

Lafi7e added training

Lafi7e requested a review from

hariharans29 4 years ago

Lafi7e requested a review from

pengwa 4 years ago

Lafi7e requested a review from

weixingzhang 4 years ago

Lafi7e requested a review from

zhijxu-MS 4 years ago

fix win build

9506df09

guyang3532 requested a review from

guyang3532 4 years ago

guyang3532 dismissed these changes on 2022-04-13

pengwa commented on 2022-04-14

pengwa dismissed these changes on 2022-04-14

rename some var

dbb6c7c2

Lafi7e dismissed their stale review via dbb6c7c2 4 years ago

pengwa dismissed these changes on 2022-04-14

Merge branch 'master' into weicwang/tile_perf

91a36632

hariharans29 commented on 2022-04-14

hariharans29 requested changes on 2022-04-14

fix typo

0dcf692c

Lafi7e dismissed their stale review via 0dcf692c 4 years ago

guyang3532 approved these changes on 2022-04-15

hariharans29 approved these changes on 2022-04-15

Lafi7e merged 0bad5b1b into master 4 years ago

Lafi7e deleted the weicwang/tile_perf branch 4 years ago

Reviewers

hariharans29

guyang3532

pengwa

weixingzhang

zhijxu-MS

Assignees

No one assigned

Labels

training

Milestone

No milestone

onnxruntime [CUDA] Rollback TileMemcpy and TileBatchedMemcpy when Block Size is Small #11187 Merged

[CUDA] Rollback TileMemcpy and TileBatchedMemcpy when Block Size is Small #11187

onnxruntime
[CUDA] Rollback TileMemcpy and TileBatchedMemcpy when Block Size is Small
#11187

Merged