Skip to content

[static runtime] Add _out variants and reuse memory #44128

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Closed
bwasti wants to merge 2 commits into from
Closed

[static runtime] Add _out variants and reuse memory #44128

bwasti wants to merge 2 commits into from

Conversation

@bwasti
Copy link
Contributor

@bwasti bwasti commented Sep 3, 2020 •
edited
Loading

Stack from ghstack:

Differential Revision: D23604304

@bwasti bwasti requested a review from apaszke as a code owner September 3, 2020 18:30
This was referenced Sep 3, 2020
Closed
Closed
Closed
Closed
@facebook-github-bot facebook-github-bot added the oncall: jit Add this issue/PR to JIT oncall triage queue label Sep 3, 2020
@dr-ci
Copy link

dr-ci bot commented Sep 3, 2020 •
edited
Loading

💊 CI failures summary and remediations

As of commit c87604a (more details on the Dr. CI page):

  • 2/2 failures possibly* introduced in this PR
    • 1/2 non-CircleCI failure(s)

🕵️ 1 new failure recognized by patterns

The following CI failures do not appear to be due to upstream breakages:

See CircleCI build pytorch_linux_backward_compatibility_check_test (1/1)

Step: "Run tests" (full log | diagnosis details | 🔁 rerun)

Sep 09 18:39:50 The PR is introducing backward incompatible changes to the operator library. Please contact PyTorch team to confirm whether this change is wanted or not. 
Sep 09 18:39:50 processing existing schema:  add_constant(__torch__.torch.classes._TorchScriptTesting._ElementwiseInterpreter _0, str _1, Tensor _2) -> (None _0) 
Sep 09 18:39:50 processing existing schema:  set_input_names(__torch__.torch.classes._TorchScriptTesting._ElementwiseInterpreter _0, str[] _1) -> (None _0) 
Sep 09 18:39:50 processing existing schema:  set_output_name(__torch__.torch.classes._TorchScriptTesting._ElementwiseInterpreter _0, str _1) -> (None _0) 
Sep 09 18:39:50 processing existing schema:  __call__(__torch__.torch.classes._TorchScriptTesting._ElementwiseInterpreter _0, Tensor[] _1) -> (Tensor _0) 
Sep 09 18:39:50 processing existing schema:  __getstate__(__torch__.torch.classes._TorchScriptTesting._ElementwiseInterpreter _0) -> ((str[], str?, Dict(str, Tensor), (str, str[], str)[]) _0) 
Sep 09 18:39:50 processing existing schema:  __setstate__(__torch__.torch.classes._TorchScriptTesting._ElementwiseInterpreter _0, (str[], str?, Dict(str, Tensor), (str, str[], str)[]) _1) -> (None _0) 
Sep 09 18:39:50 processing existing schema:  get(__torch__.torch.classes._TorchScriptTesting._LiteInterpreterTest _0, Tensor _1) -> (str _0) 
Sep 09 18:39:50 processing existing schema:  __getstate__(__torch__.torch.classes._TorchScriptTesting._LiteInterpreterTest _0) -> (int _0) 
Sep 09 18:39:50 processing existing schema:  __setstate__(__torch__.torch.classes._TorchScriptTesting._LiteInterpreterTest _0, int _1) -> (None _0) 
Sep 09 18:39:50 processing existing schema:  __init__(__torch__.torch.classes.dist_rpc.WorkerInfo _0, str _1, int _2) -> (None _0) 
Sep 09 18:39:50 The PR is introducing backward incompatible changes to the operator library. Please contact PyTorch team to confirm whether this change is wanted or not.  
Sep 09 18:39:50  
Sep 09 18:39:50 Broken ops: [ 
Sep 09 18:39:50 	static::add(Tensor a, Tensor b) -> (Tensor) 
Sep 09 18:39:50 	static::mul.a(Tensor a, Tensor b) -> (Tensor) 
Sep 09 18:39:50 	static::mul.b(Tensor a, int b) -> (Tensor) 
Sep 09 18:39:50 ] 
Sep 09 18:39:50 + cleanup 
Sep 09 18:39:50 + retcode=1 
Sep 09 18:39:50 + set +x 
Sep 09 18:39:50 =================== sccache compilation log ===================

ci.pytorch.org: 1 failed

This comment was automatically generated by Dr. CI (expand for details).Follow this link to opt-out of these comments for your Pull Requests.

Please report bugs/suggestions on the GitHub issue tracker or post in the (internal) Dr. CI Users group.

See how this bot performed.

This comment has been revised 7 times.

@bwasti bwasti changed the title Add _out variants and reuse memory [static runtime] Add _out variants and reuse memory Sep 4, 2020
hlu1
hlu1 approved these changes Sep 9, 2020
View reviewed changes
torch/csrc/jit/runtime/static/ops.cpp
Comment on lines +9 to +13
if ((str == "aten::add") || (str == "aten::mul") || (str == "aten::addmm") ||
(str == "aten::bmm") || (str == "aten::sigmoid") ||
(str == "aten::cat")) {
return true;
}
Copy link
Contributor

@hlu1 hlu1 Sep 9, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think it's better to keep the op names in a set, so the lookup time is O(1). We'll add more and more ops to the out_of_place set so eventually we can get full coverage of the models we support.

@facebook-github-bot
Copy link
Contributor

facebook-github-bot commented Sep 25, 2020

@bwasti merged this pull request in d1a1161.

@facebook-github-bot facebook-github-bot deleted the gh/bwasti/47/head branch September 29, 2020 14:23
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@apaszke apaszke Awaiting requested review from apaszke

1 more reviewer

@hlu1 hlu1 hlu1 approved these changes

Reviewers whose approvals may not affect merge requirements

Assignees

No one assigned

Labels

Merged oncall: jit Add this issue/PR to JIT oncall triage queue

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

5 participants

@bwasti @facebook-github-bot @hlu1 @mruberry