-
Notifications
You must be signed in to change notification settings - Fork 16
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.
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
Optimize greedy rewrite passes #298
Optimize greedy rewrite passes #298
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
config.strictMode = mlir::GreedyRewriteStrictness::ExistingOps; | ||
// Each operation can only be modified once so limit | ||
config.maxIterations = 1; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Of these two changes, which provides more benefit? I'm surprised that the first changes things at all since the rewrite pattern isn't adding any new operations.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think its maxIterations
bee5330
to
be9f222
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
Optimize a number of greedy rewrite passes. - QUIRAngleConversionPass: Goes from about ~7.5s for 100x100 to 0.9s on my machine
Optimize a number of greedy rewrite passes. - QUIRAngleConversionPass: Goes from about ~7.5s for 100x100 to 0.9s on my machine
Optimize a number of greedy rewrite passes.