[PATCH] D23388: [LoopUnroll] By default disable unrolling when optimizing for size.

[Michael Zolotukhin via Phabricator via llvm-commits]

mzolotukhin added a comment.

> I looked at your benchmarks numbers, with particular attention to the worst regressions, and I'm not convinced these benchmarks are representative.

What about tramp3d-v4 and SPECs? As far as I remember, tramp3d-v4 compiles pretty long and is pretty stable (and it is a part of CTMark).

But anyway, I completely agree with:

> I think we should try to base our decisions (at least for compile time choices on more real-world/CPU intensive programs/workloads).

I'll be happy to look at patches for tuning the Os thresholds! If noone picks it up, I might return to it at some point, but it's not in my near-term plans now.

Thanks,
Michael


Repository:
  rL LLVM

https://reviews.llvm.org/D23388