[PATCH] D85460: [SVE] No need to AND(X, 0x1) when truncating a scalable mask

[Cameron McInally via Phabricator via llvm-commits]

cameron.mcinally created this revision.
cameron.mcinally added reviewers: paulwalker-arm, sdesmalen, efriedma, rengolin.
Herald added subscribers: llvm-commits, psnobl, hiraditya, tschuett.
Herald added a project: LLVM.
cameron.mcinally requested review of this revision.

We're using CMPNE(X, 0) to truncate promoted fixed width masks to scalable masks. My confidence is moderate on this, but I think the preceding AND(X, 0x1) is unnecessary. I wouldn't expect a false element with tops bits set.

Did I miss something??


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D85460

Files:
  llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
  llvm/test/CodeGen/AArch64/sve-trunc.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D85460.283683.patch
Type: text/x-patch
Size: 3851 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200806/22392ae3/attachment.bin>