OSchip
/
llvm-project
mirror of https://github.com/THU-DSP-LAB/llvm-project.git
13
0
Fork 6
Code Issues Packages Projects Releases Wiki Activity
242,203 Commits 32 Branches 0 Tags 2.2 GiB
Commit Graph

5 Commits

Author SHA1 Message Date
Sanjay Patel 2d477e59e8 [InstCombine] fold add(zext(xor X, C), C) --> sext X when C is INT_MIN in the source type 
The pattern may look more obviously like a sext if written as:

  define i32 @g(i16 %x) {
    %zext = zext i16 %x to i32
    %xor = xor i32 %zext, 32768
    %add = add i32 %xor, -32768
    ret i32 %add
  }

We already have that fold in visitAdd().

Differential Revision: https://reviews.llvm.org/D22477

llvm-svn: 276035
 2016-07-19 22:09:34 +00:00
Sanjay Patel dbf44f5016 add tests for missed sext transform 
llvm-svn: 275908
 2016-07-18 20:37:51 +00:00
Sanjay Patel 79acd2a96b [InstCombine] allow X + signbit --> X ^ signbit for vector splats 
llvm-svn: 275691
 2016-07-16 18:29:26 +00:00
Sanjay Patel 040bd16e56 add vector test to show missing transform 
llvm-svn: 275690
 2016-07-16 18:24:18 +00:00
Sanjay Patel eb50476f77 update tests to use FileCheck, consolidate tests, fix comments 
llvm-svn: 275688
 2016-07-16 18:08:22 +00:00