Add initial support for PGO instrumentation/use in inputgen #2
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
This patch adds flags (and a test) so that the main input-gen binary can add instrumentation to modules and also compile the run modules with a PGO profile. This enables downstream uses of PGO, such as block coverage analysis. Passing the flags directly into the input-gen binary for recompiling with a PGO profile probably isn't what will be used for large scale testing (instead just opting to recompile the bitcode), but probably won't hurt to support. This is also currently not selective in terms of what gets instrumented. This means that parts of the runtime will end up in the profile. This shouldn't be an issue, but something that we may want to work on in the future.
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Outdated suggestions cannot be applied.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.
This patch adds flags (and a test) so that the main input-gen binary can add instrumentation to modules and also compile the run modules with a PGO profile. This enables downstream uses of PGO, such as block coverage analysis. Passing the flags directly into the input-gen binary for recompiling with a PGO profile probably isn't what will be used for large scale testing (instead just opting to recompile the bitcode), but probably won't hurt to support.
This is also currently not selective in terms of what gets instrumented. This means that parts of the runtime will end up in the profile. This shouldn't be an issue, but something that we may want to work on in the future.