Trying to use WhiteFox on LLVM optimizations

[zitongzhoueric]

Hi I'm trying to use WhiteFox on LLVM optimizations.

Please correct me if I'm wrong, I don't seem to see the code that executes LLVM and refines prompts using its feedback? Can you provide some instructions on how to run Whitefox with feedback on LLVM?
My understanding is that python starcoder_gen.py --prompt-dir=Prompts/llvm/req2test --output-dir=starcoder-generated-llvm --num=... is one-shot and doesn't use any feedback?

Also there are only prompts for 21 LLVM optimizations in WhiteFox/Prompts/llvm/req2test. Is it possible to include the prompts for other Transform passes in https://llvm.org/docs/Passes.html that you have tested?

Thank you!

[YangChenyuan]

Thanks for your interest in WhiteFox!

Can you provide some instructions on how to run Whitefox with feedback on LLVM?

Sure! First, you need to instrument the LLVM. I will upload the patch later.

In the meantime, you can use the pre-generated one-shot results from the req2test directory as the WhiteFox output for LLVM.

Also there are only prompts for 21 LLVM optimizations in WhiteFox/Prompts/llvm/req2test

I just uploaded all prompts for LLVM by PR #11

If you have any further questions, feel free to ask me!