maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

AMDGPU/GlobalISel: Remove dead patterns 

llvm-svn: 371404
This commit is contained in:
Matt Arsenault 2019-09-09 15:06:06 +00:00
parent 783fc95f3e
commit 508dff2ce1
1 changed files with 0 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
llvm/lib/Target/AMDGPU/AMDGPUGISel.td
View File

@ -173,11 +173,6 @@ def : GISelVop2Pat <or, V_OR_B32_e32, i32>;
def : GISelVop3Pat2ModsPat <
int_amdgcn_cvt_pkrtz, V_CVT_PKRTZ_F16_F32_e64, v2f16, f32>;
defm : GISelVop2IntrPat <int_maxnum, V_MAX_F32_e32, f32>;
def : GISelVop3Pat2ModsPat <int_maxnum, V_MAX_F64, f64>;
defm : GISelVop2IntrPat <int_minnum, V_MIN_F32_e32, f32>;
def : GISelVop3Pat2ModsPat <int_minnum, V_MIN_F64, f64>;
// Since GlobalISel is more flexible then SelectionDAG, I think we can get
// away with adding patterns for integer types and not legalizing all
// loads and stores to vector types. This should help simplify the load/store