News for hackers. Stuff that matters.

A comment on Souper – A Superoptimizer for LLVM IR. Parent comment

Re:Souper – A Superoptimizer for LLVM IR by cwzwarich on Wednesday October 28, 2015 @07:57AM (#10463892)
It's been done:

http://theory.stanford.edu/~aiken/publications/papers/asplos...

They had to restrict themselves to instruction sequences of length 3 due to limited resources, and say that to go beyond instruction sequences of length 4 they would need a better approach.

[ Parent | on HN ]
Re:Souper – A Superoptimizer for LLVM IR by fulafel on Wednesday October 28, 2015 @08:02AM (#10463907)
This has been done in '87 by Massalin: https://www.cs.arizona.edu/~collberg/Teaching/553/2011/Resou...

(both are brute force though, not solver-based)

[ Parent | on HN ]
Re:Souper – A Superoptimizer for LLVM IR by cwzwarich on Wednesday October 28, 2015 @08:16AM (#10463960)
Massalin's superoptimizer required humans to check the correctness of optimizations. It also didn't generate parameterized peephole optimizations; it directly optimized concrete instruction sequences.
[ Parent | on HN ]