Commit Graph

8 Commits

Author SHA1 Message Date
Philip Reames 449d14ebd2 Do actual DCE in LoopUnroll (try 4)
Turns out simplifyLoopIVs sometimes returns a non-dead instruction in it's DeadInsts out param.  I had done a bit of NFC cleanup which was only NFC if simplifyLoopIVs obeyed it's documentation.  I'm simplfy dropping that part of the change.

Commit message from try 3:

Recommitting after fixing a bug found post commit. Amusingly, try 1 had been correct, and by reverting to incorporate last minute review feedback, I introduce the bug. Oops. :)

Original commit message:

The problem was that recursively deleting an instruction can delete instructions beyond the current iterator (via a dead phi), thus invalidating iteration. Test case added in LoopUnroll/dce.ll to cover this case.

LoopUnroll does a limited DCE pass after unrolling, but if you have a chain of dead instructions, it only deletes the last one. Improve the code to recursively delete all trivially dead instructions.

Differential Revision: https://reviews.llvm.org/D102511
2021-05-19 10:25:31 -07:00
Amy Huang 517857421d Revert "Do actual DCE in LoopUnroll (try 3)"
This reverts commit b6320eeb86
as it causes clang to assert; see
https://reviews.llvm.org/rGb6320eeb8622f05e4a5d4c7f5420523357490fca.
2021-05-19 08:53:38 -07:00
Philip Reames b6320eeb86 Do actual DCE in LoopUnroll (try 3)
Recommitting after fixing a bug found post commit.  Amusingly, try 1 had been correct, and by reverting to incorporate last minute review feedback, I introduce the bug.  Oops.  :)

The problem was that recursively deleting an instruction can delete instructions beyond the current iterator (via a dead phi), thus invalidating iteration.  Test case added in LoopUnroll/dce.ll to cover this case.

LoopUnroll does a limited DCE pass after unrolling, but if you have a chain of dead instructions, it only deletes the last one. Improve the code to recursively delete all trivially dead instructions.

Differential Revision: https://reviews.llvm.org/D102511
2021-05-17 14:47:02 -07:00
Philip Reames 6ae9893ed2 Revert "Do actual DCE in LoopUnroll (try 2)"
This reverts commit 653fa0b46a.

Reported to trigger pr50354.  Reverting until investigated.
2021-05-16 09:38:36 -07:00
Philip Reames 653fa0b46a Do actual DCE in LoopUnroll (try 2)
Recommitting after addressing a missed review comment, and updating an aarch64 test I'd missed.

LoopUnroll does a limited DCE pass after unrolling, but if you have a chain of dead instructions, it only deletes the last one. Improve the code to recursively delete all trivially dead instructions.

Differential Revision: https://reviews.llvm.org/D102511
2021-05-14 10:42:36 -07:00
Philip Reames e488bf815f Revert "Do actual DCE in LoopUnroll"
This reverts commit 9d1a61e695.

I'd missed some review feedback, and had missed updating an aarch64 test.  Reverting while I fix both.
2021-05-14 10:15:30 -07:00
Philip Reames 9d1a61e695 Do actual DCE in LoopUnroll
LoopUnroll does a limited DCE pass after unrolling, but if you have a chain of dead instructions, it only deletes the last one. Improve the code to recursively delete all trivially dead instructions.

Differential Revision: https://reviews.llvm.org/D102511
2021-05-14 10:05:25 -07:00
Florian Hahn 1bd58870e5 [LoopUnroll] Use LoopSize+1 as threshold, to allow unrolling loops matching LoopSize.
We use `< UP.Threshold` later on, so we should use LoopSize + 1, to
allow unrolling if the result won't exceed to loop size.

Fixes PR43305.

Reviewers: efriedma, dmgreen, paquette

Reviewed By: dmgreen

Differential Revision: https://reviews.llvm.org/D67594

llvm-svn: 372084
2019-09-17 09:02:48 +00:00