nikic wrote: > but a question for @dtcxzyw and @nikic I assume that we can not solve all > regressions so when is it good enough?
I guess basically when there are more improvements than regressions, and the known regressions can't be easily fixed? I briefly scrolled through the llvm-opt-benchmark results, and I think there are still some "simple" regressions there. For example there's this missed optimization: https://alive2.llvm.org/ce/z/kWLxGe https://github.com/llvm/llvm-project/pull/184182 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
