onnxruntime/orttraining
Weixing Zhang ca9b3f18e9
Explicitly pass cuda stream to thrust function rather than use cuda default stream implicitly (#7414)
* Pass cuda stream to thrust function to not use default stream.

In the commit 299ace0, ORT has been changed to not use cuda default stream.

* update amd_hipify.py

* remove un-necessary stream sync

Co-authored-by: Weixing Zhang <wezhan@microsoft.com>
2021-04-25 01:18:56 -07:00
..
orttraining Explicitly pass cuda stream to thrust function rather than use cuda default stream implicitly (#7414) 2021-04-25 01:18:56 -07:00
pytorch_frontend_examples Sync ORTModule branch with master and fix tests (#6526) 2021-02-02 08:59:56 -08:00
tools Add BERT-L perf regression test on MI100 and re-enable batch size test (#7240) 2021-04-05 15:51:52 -07:00