[PATCH] D43973: [AArch64] define isExtractSubvectorCheap

Sebastian Pop via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Thu Mar 1 14:30:06 PST 2018


sebpop created this revision.
sebpop added reviewers: efriedma, SjoerdMeijer.
Herald added subscribers: hiraditya, kristof.beyls, javed.absar, rengolin.

Following the ARM-neon backend, define isExtractSubvectorCheap to return true
when extracting low and high part of a neon register.

      

The patch disables a test in llvm/test/CodeGen/AArch64/arm64-ext.ll that was
checking that ReconstructShuffle in ISelLowering is working as expected. The
pattern to exercise ReconstructShuffle is a BUILD_VECTOR and the expected
pattern gets transformed earlier by the DAGCombiner into an extract_subvector +
vector_shuffle. As there is no way to disable the combiner to only exercise the
code in ISelLowering, the patch disables the testcase.


Repository:
  rL LLVM

https://reviews.llvm.org/D43973

Files:
  llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
  llvm/lib/Target/AArch64/AArch64ISelLowering.h
  llvm/test/CodeGen/AArch64/aarch64-vuzp.ll
  llvm/test/CodeGen/AArch64/arm64-ext.ll
  llvm/test/CodeGen/AArch64/neon-scalar-copy.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D43973.136604.patch
Type: text/x-patch
Size: 5656 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20180301/4d23ef77/attachment.bin>


More information about the llvm-commits mailing list