alokkrsharma wrote:

> I can't speak for the PGOVerify code itself, because PGO isn't my strong 
> suit, but the code in StandardInstrumentations.h looks good to me. However, 
> with so much of the patch left unreviewed by me, I don't feel like I can 
> approve it either. Please note the merge conflicts that need to be fixed up

Thanks for taking a look, I really appreciate the review.

It's helpful to know that the pass-instrumentation integration looks 
reasonable, even if the PGO-specific parts are outside your area of expertise.

Just to clarify, I'm not proposing this PR for immediate landing. The branch is 
primarily intended as a proof-of-concept to support the RFC discussion, 
validate the design on real pipelines, and evaluate the signal/noise 
characteristics of the verifier.

I'll fix up the merge conflicts as well.

Once the RFC direction becomes clearer, I expect the work would be split into 
smaller reviewable pieces rather than upstreamed as a single large change.

https://github.com/llvm/llvm-project/pull/196295
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to