Add torch.distributed get_rank and get_world_size to constant_fold_functions #109029
Conversation
…nctions Signed-off-by: Edward Z. Yang <ezyang@meta.com> [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/109029
Note: Links to docs will display an error until the docs builds have been completed. ✅ You can merge normally! (4 Unrelated Failures)As of commit aa7a84d with merge base 090fe45 ( FLAKY - The following job failed but was likely due to flakiness present on trunk:
UNSTABLE - The following jobs failed but were likely due to flakiness present on trunk and has been marked as unstable:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
github-actions
bot
requested review from
SherlockNoMad,
albanD,
antoniojkim,
bdhirsh,
miladm,
voznesenskym and
wconstab
torch/_dynamo/variables/torch.py
Outdated
| torch.nn.functional._Reduction.get_enum, | ||
| torch._C._get_privateuse1_backend_name, | ||
| torch.distributed.get_rank, | ||
| torch.distributed.get_world_size, |
There was a problem hiding this comment.
basic Q: it seems a bit redundant to force dynamo to inline a hardcoded list of torch API's into the graph, if we know that AOTAutograd will do it later anyway. Is main purpose to (mildly) limit the API surface for any backends that want to do stuff before AOTAutograd?
There was a problem hiding this comment.
It's actually not inlining; this says that it's OK to run the operation, get out a Python constant, and treat it as constant for the purposes of the rest of Dynamo execution
…ant_fold_functions" Signed-off-by: Edward Z. Yang <ezyangmeta.com> cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng Xia-Weiwen wenzhe-nrv jiayisunx chenyang78 aakhundov kadeng [ghstack-poisoned]
|
@pytorchbot merge |
Merge failedReason: This PR needs a If not, please add the To add a label, you can comment to pytorchbot, for example For more information, see Details for Dev Infra teamRaised by workflow job |
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Merge failedReason: 1 jobs have failed, first few of them are: trunk / macos-12-py3-arm64-mps / test (default, 1, 1) Details for Dev Infra teamRaised by workflow job |
|
@pytorchbot merge -i |
Merge startedYour change will be merged while ignoring the following 4 checks: pull / linux-focal-py3.8-clang10 / test (dynamo, 1, 2, linux.2xlarge, unstable), pull / linux-focal-py3.11-clang10 / test (dynamo, 2, 2, linux.2xlarge, unstable), inductor / cuda12.1-py3.10-gcc9-sm86 / test (inductor, 1, 1, linux.g5.4xlarge.nvidia.gpu, unstable), trunk / macos-12-py3-arm64-mps / test (default, 1, 1) Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
3 participants
Stack from ghstack (oldest at bottom):
Signed-off-by: Edward Z. Yang ezyang@meta.com
cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @Xia-Weiwen @wenzhe-nrv @jiayisunx @chenyang78 @aakhundov @kadeng