[PATCH] D30562: Add Instruction number to LSR cost model (PR23384) part 3 of 3

[Evgeny Stupachenko via Phabricator via llvm-commits]

evstupac created this revision.

The patch depends on https://reviews.llvm.org/D30561 (compilation will fail without it).
It makes Insns number priority number in LSR cost for X86
and
Adds Cross Uses to make possible solution without cmp  (less on 1 instruction) - this is required by some tests to be consistent (choose solution without cmp for all cases).


Repository:
  rL LLVM

https://reviews.llvm.org/D30562

Files:
  lib/Target/X86/X86TargetTransformInfo.cpp
  lib/Target/X86/X86TargetTransformInfo.h
  lib/Transforms/Scalar/LoopStrengthReduce.cpp
  test/CodeGen/X86/2006-05-11-InstrSched.ll
  test/CodeGen/X86/atom-fixup-lea3.ll
  test/CodeGen/X86/avoid_complex_am.ll
  test/CodeGen/X86/compact-unwind.ll
  test/CodeGen/X86/full-lsr.ll
  test/CodeGen/X86/loop-strength-reduce4.ll
  test/CodeGen/X86/masked-iv-safe.ll
  test/CodeGen/X86/misched-matrix.ll
  test/Transforms/LoopStrengthReduce/X86/ivchain-X86.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D30562.90426.patch
Type: text/x-patch
Size: 16771 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170303/3df2d6f2/attachment.bin>