[all-commits] [llvm/llvm-project] d0907c: [LoopUnroll] Directly update DT instead of DTU.

[Florian Hahn via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: d0907ce7ed9f159562ca3f4cfd8d87e89e93febe
      https://github.com/llvm/llvm-project/commit/d0907ce7ed9f159562ca3f4cfd8d87e89e93febe
  Author: Florian Hahn <flo at fhahn.com>
  Date:   2023-01-19 (Thu, 19 Jan 2023)

  Changed paths:
    M llvm/include/llvm/Transforms/Utils/BasicBlockUtils.h
    M llvm/lib/Transforms/Utils/BasicBlockUtils.cpp
    M llvm/lib/Transforms/Utils/LoopUnroll.cpp

  Log Message:
  -----------
  [LoopUnroll] Directly update DT instead of DTU.

The scope of DT updates are very limited when unrolling loops: the DT
should only need updating for
* new blocks added
* exiting blocks we simplified branches

This can be done manually without too much extra work.
MergeBlockIntoPredecessor also needs to be updated to support direct
DT updates.

This fixes excessive time spent in DTU for same cases. In an internal
example, time spent in LoopUnroll with this patch goes from ~200s to 2s.

It also is slightly positive for CTMark:
* NewPM-O3: -0.13%
* NewPM-ReleaseThinLTO: -0.11%
* NewPM-ReleaseLTO-g: -0.13%

Notable improvements are mafft (~ -0.50%) and lencod (~ -0.30%), with no
workload regressed.

https://llvm-compile-time-tracker.com/compare.php?from=78a9ee7834331fb4360457cc565fa36f5452f7e0&to=687e08d011b0dc6d3edd223612761e44225c7537&stat=instructions:u

Reviewed By: kuhar

Differential Revision: https://reviews.llvm.org/D141487