search for: canical

2015-07-08 17:58 GMT+02:00 escha <escha at apple.com>: > [...] > > If you want to “revert" this sort of thing, you can do it at Select() time > or PreprocessISelDAG(), which is what I did on an out-of-tree backend to > turn add X, -C into sub X, C on selection time. This still lets all the > intermediate optimizations take advantage of the canonicalization. > >

...t we do not need to spend time rebuilding them later. > Week 8-9: Rewrite some > expensive functions for Polly code generation based on profiling > information. This is still very vague. I propose to > StageIV -- Let Polly bail out early. [Week 10] > > Week 10: Add support in canicalization step or optimization step to Typo -----> canonicalization > allow Polly boil out early if it cannot benefit programs. > StageV -- Improve other parts. [Week 11-12] > > Week 11: Improve other parts of Polly. Especially, I will focus on some > expensive help...

Hi all, I have updated my GSoS proposal: "FastPolly: Reducing LLVM-Polly Compiling overhead" (https://gist.github.com/tanstar/5441808). I think the pass ordering problem you discussed early can be also investigated in this project! Is there any comment or advice about my proposal? I appreciate all your help and advice. Thanks, Star Tan Proposal: