[PyTorch][JIT] Don't refcount Type singletons#69579
Closed
swolchok wants to merge 13 commits into
Closed
Conversation
This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/) [ghstack-poisoned]
Contributor
🔗 Helpful links
💊 CI failures summary and remediationsAs of commit 5eb65da (more details on the Dr. CI page): 💚 💚 Looks good so far! There are no failures yet. 💚 💚 This comment was automatically generated by Dr. CI (expand for details).Please report bugs/suggestions to the (internal) Dr. CI Users group. |
facebook-github-bot
added
the
oncall: jit
CI Flow Status⚛️ CI FlowRuleset - Version:
You can add a comment to the PR and tag @pytorchbot with the following commands: # ciflow rerun, "ciflow/default" will always be added automatically
@pytorchbot ciflow rerun
# ciflow rerun with additional labels "-l <ciflow/label_name>", which is equivalent to adding these labels manually and trigger the rerun
@pytorchbot ciflow rerun -l ciflow/scheduled -l ciflow/slowFor more information, please take a look at the CI Flow Wiki. |
This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/) [ghstack-poisoned]
swolchok
added a commit
that referenced
this pull request
Dec 8, 2021
Pull Request resolved: #69579 This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. ghstack-source-id: 145056617 Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/)
This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/) [ghstack-poisoned]
swolchok
added a commit
that referenced
this pull request
Dec 8, 2021
Pull Request resolved: #69579 This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. ghstack-source-id: 145145357 Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/)
Member
|
Should I review? Tests look sadge |
Contributor
Author
you can still review for broad direction. blocking on tests slows down the process. |
Contributor
Author
|
Looks like there's some kind of CUDA build failure I don't understand (can we not do explicit destructor calls in CUDA? why is this code getting built for CUDA anyway?), and I have to patch the Python bridging stuff for the new classes. |
This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/) [ghstack-poisoned]
swolchok
added a commit
that referenced
this pull request
Dec 14, 2021
Pull Request resolved: #69579 This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. ghstack-source-id: 145671620 Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/)
This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/) [ghstack-poisoned]
swolchok
added a commit
that referenced
this pull request
Dec 15, 2021
Pull Request resolved: #69579 This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. ghstack-source-id: 145771095 Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/)
This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/) [ghstack-poisoned]
This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/) [ghstack-poisoned]
swolchok
added a commit
that referenced
this pull request
Dec 16, 2021
Pull Request resolved: #69579 This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. ghstack-source-id: 145793691 Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/)
This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/) [ghstack-poisoned]
swolchok
added a commit
that referenced
this pull request
Dec 16, 2021
Pull Request resolved: #69579 This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. ghstack-source-id: 145813052 Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/)
swolchok
commented
Dec 16, 2021
Comment on lines
+236
to
+241
| // Without SharedPtrWrapper, this line would read | ||
| // `shared_.~shared_ptr()` and nvcc would complain with | ||
| // "error: expected primary-expression before '>' token" | ||
| // referring to the "t" in "shared_ptr". SharedPtrWrapper | ||
| // exists to work around this compiler bug. | ||
| shared_.~SharedPtrWrapper(); |
This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/) [ghstack-poisoned]
This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/) [ghstack-poisoned]
swolchok
added a commit
that referenced
this pull request
Dec 16, 2021
Pull Request resolved: #69579 This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. ghstack-source-id: 145875239 Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D32923880/)!
suo
approved these changes
Dec 20, 2021
This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/) [ghstack-poisoned]
swolchok
added a commit
that referenced
this pull request
Jan 4, 2022
Pull Request resolved: #69579 This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. ghstack-source-id: 146468426 Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D32923880/)!
This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/) [ghstack-poisoned]
swolchok
added a commit
that referenced
this pull request
Jan 5, 2022
Pull Request resolved: #69579 This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. ghstack-source-id: 146558363 Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D32923880/)!
This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/) [ghstack-poisoned]
swolchok
added a commit
that referenced
this pull request
Jan 6, 2022
Pull Request resolved: #69579 This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. ghstack-source-id: 146643993 Differential Revision: [D32923880](https://our.internmc.facebook.com/intern/diff/D32923880/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D32923880/)!
Contributor
Author
|
I've conservatively labeled this bc-breaking because it is possible for it to cause compile errors in downstream code that tried to use std::dynamic_pointer_cast on TypePtrs instead of going through the Type::cast API. I ran into one example of this internally and the fix was to just use Type::cast. |
laurentdupin
pushed a commit
to laurentdupin/pytorch
that referenced
this pull request
Apr 25, 2026
Summary: Pull Request resolved: pytorch#69579 This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem. ghstack-source-id: 146643993 Test Plan: Ran //caffe2/caffe2/fb/high_perf_models/pytorch/benchmark_framework_overheads:cpp_benchmark with arguments `--op empty -niter 40 --stressTestRecordFunction --captureRecordFunctionInputs` on devbig with turbo off. Before: ``` I1206 13:47:15.037441 1201670 bench.cpp:144] Mean 0.737675 I1206 13:47:15.037463 1201670 bench.cpp:145] Median 0.736725 I1206 13:47:15.037468 1201670 bench.cpp:146] Min 0.722897 I1206 13:47:15.037473 1201670 bench.cpp:147] stddev 0.00508187 I1206 13:47:15.037482 1201670 bench.cpp:148] stddev / mean 0.00688903 ``` After: ``` I1206 13:48:16.830123 1205612 bench.cpp:144] Mean 0.66988 I1206 13:48:16.830150 1205612 bench.cpp:145] Median 0.663956 I1206 13:48:16.830157 1205612 bench.cpp:146] Min 0.65986 I1206 13:48:16.830164 1205612 bench.cpp:147] stddev 0.0335928 I1206 13:48:16.830171 1205612 bench.cpp:148] stddev / mean 0.0501475 ``` Static runtime startup is also improved; for CMF local_ro, time to initialize a predictor went from 10.01s to 9.59s. (Note: I wish I had a production workload to demonstrate the advantage of this on. I tried ctr_mobile_feed local_ro net but it was neutral. Anything that manipulates types or List/Dict a lot might be promising.) Reviewed By: suo Differential Revision: D32923880 fbshipit-source-id: c82ed6689b3598e61047fbcb2149982173127ff0
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Stack from ghstack:
This should help us avoid reference counting overhead on singleton Type subclasses without a major rewrite of the Type subsystem.
Differential Revision: D32923880
cc @ezyang @gchanan