[jit] correctly share types between traced modules #29583
Closed
Conversation
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
The normal flow for type sharing assumes that we will construct the `ConcreteModuleType`, then use `operator==` to decide whether or not to reuse an existing JIT type. In this case, `jitType_` is not populated, so it doesn't make sense to compare it. However, there is one exception to this flow: for traced modules, we pre-compute the JIT type and poke it into the `ConcreteModuleType` manually. To handle this case, we should compare the `jitType_`s in `operator==` like everything else.
facebook-github-bot
added
the
oncall: jit
The normal flow for type sharing assumes that we will construct the `ConcreteModuleType`, then use `operator==` to decide whether or not to reuse an existing JIT type. In this case, `jitType_` is not populated, so it doesn't make sense to compare it. However, there is one exception to this flow: for traced modules, we pre-compute the JIT type and poke it into the `ConcreteModuleType` manually. To handle this case, we should compare the `jitType_`s in `operator==` like everything else. Differential Revision: [D18435949](https://our.internmc.facebook.com/intern/diff/D18435949)
suo
added a commit
that referenced
this pull request
Nov 11, 2019
The normal flow for type sharing assumes that we will construct the `ConcreteModuleType`, then use `operator==` to decide whether or not to reuse an existing JIT type. In this case, `jitType_` is not populated, so it doesn't make sense to compare it. However, there is one exception to this flow: for traced modules, we pre-compute the JIT type and poke it into the `ConcreteModuleType` manually. To handle this case, we should compare the `jitType_`s in `operator==` like everything else. ghstack-source-id: ffc21c9 Pull Request resolved: #29583
wanchaol
approved these changes
Nov 11, 2019
Member
CircleCI build failures summaryAs of commit 1c976a8:
Here are the reasons each build failed. This comment was automatically generated by Dr. CI. Please report bugs/suggestions on the GitHub issue tracker. This comment has been revised 1 time(s). |
Contributor
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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:
ConcreteModuleType::dump()#29582 ImproveConcreteModuleType::dump()The normal flow for type sharing assumes that we will construct the
ConcreteModuleType, then useoperator==to decide whether or not toreuse an existing JIT type. In this case,
jitType_is not populated,so it doesn't make sense to compare it.
However, there is one exception to this flow: for traced modules, we
pre-compute the JIT type and poke it into the
ConcreteModuleTypemanually. To handle this case, we should compare the
jitType_s inoperator==like everything else.Differential Revision: D18435949