[PATCH] D48168: AMDGPU: Remove redundant MIMG instruction variants

Nicolai Hähnle via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Thu Jun 14 06:24:42 PDT 2018


nhaehnle created this revision.
nhaehnle added reviewers: arsenm, rampitec.
Herald added subscribers: t-tye, tpr, dstuttard, yaxunl, wdng, kzhuravl.

For sample and gather ops, we can accurately determine the set of
vaddr-size instruction variants that are required. This reduces
the size of instruction tables by ~5%.

The number of machine instruction opcodes is reduced from 10002
to 9476.

Change-Id: Ie7fc65d3657b762c7816017fe70b2e9bec644a8a


Repository:
  rL LLVM

https://reviews.llvm.org/D48168

Files:
  lib/Target/AMDGPU/MIMGInstructions.td
  test/MC/AMDGPU/gfx7_asm_all.s
  test/MC/AMDGPU/gfx8_asm_all.s

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D48168.151340.patch
Type: text/x-patch
Size: 149207 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20180614/b90ec87d/attachment.bin>


More information about the llvm-commits mailing list