[PATCH] D69538: [ModuloSchedule] Fix experimental modulo expansion for data loop carried dependencies.

[Thomas Raoux via Phabricator via llvm-commits]

ThomasRaoux created this revision.
ThomasRaoux added a reviewer: jmolloy.
Herald added subscribers: llvm-commits, hiraditya.
Herald added a project: LLVM.

The new experimental expansion has a problem when a value has a data dependency with am instruction from a previous stage. This is due to the way we peel out the kernel. 
To fix that I'm changing the way we peel out the kernel. We now peel the kernel NumberStage - 1 times. The code would be correct at this point if we didn't have to handle cases where the loop iteration is smaller than the number of stages. To handle this case we move instructions between different epilogues based on their stage and remap the PHI instructions correctly.


https://reviews.llvm.org/D69538

Files:
  llvm/include/llvm/CodeGen/ModuloSchedule.h
  llvm/lib/CodeGen/ModuloSchedule.cpp
  llvm/test/CodeGen/Hexagon/swp-conv3x3-nested.ll
  llvm/test/CodeGen/Hexagon/swp-epilog-phi12.ll
  llvm/test/CodeGen/Hexagon/swp-stages4.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D69538.226777.patch
Type: text/x-patch
Size: 12127 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20191028/c607c818/attachment.bin>