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

[AMDGPU] Add some missing always-uniform values. 

This commit adds some missing intrinsics into the isAlwaysUniform list
for the AMDGPU backend.

Differential Revision: https://reviews.llvm.org/D56845

llvm-svn: 351562
This commit is contained in:
Neil Henning 2019-01-18 16:39:27 +00:00
parent 7503316d10
commit 3ed09f8e0c
2 changed files with 21 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp
View File

@ -579,6 +579,8 @@ bool GCNTTIImpl::isAlwaysUniform(const Value *V) const {
return false;
case Intrinsic::amdgcn_readfirstlane:
case Intrinsic::amdgcn_readlane:
case Intrinsic::amdgcn_icmp:
case Intrinsic::amdgcn_fcmp:
return true;
}
}

20
llvm/test/Analysis/DivergenceAnalysis/AMDGPU/always_uniform.ll
View File

@ -1,6 +1,7 @@
; RUN: opt -mtriple amdgcn-unknown-amdhsa -analyze -divergence -use-gpu-divergence-analysis %s | FileCheck %s
define amdgpu_kernel void @workitem_id_x() #1 {
; CHECK: for function 'readfirstlane':
define amdgpu_kernel void @readfirstlane() {
%id.x = call i32 @llvm.amdgcn.workitem.id.x()
; CHECK: DIVERGENT: %id.x = call i32 @llvm.amdgcn.workitem.id.x()
%first.lane = call i32 @llvm.amdgcn.readfirstlane(i32 %id.x)
@ -8,7 +9,24 @@ define amdgpu_kernel void @workitem_id_x() #1 {
ret void
}
; CHECK: for function 'icmp':
define amdgpu_kernel void @icmp(i32 inreg %x) {
; CHECK-NOT: DIVERGENT: %icmp = call i64 @llvm.amdgcn.icmp.i32
%icmp = call i64 @llvm.amdgcn.icmp.i32(i32 %x, i32 0, i32 33)
ret void
}
; CHECK: for function 'fcmp':
define amdgpu_kernel void @fcmp(float inreg %x, float inreg %y) {
; CHECK-NOT: DIVERGENT: %fcmp = call i64 @llvm.amdgcn.fcmp.i32
%fcmp = call i64 @llvm.amdgcn.fcmp.i32(float %x, float %y, i32 33)
ret void
}
declare i32 @llvm.amdgcn.workitem.id.x() #0
declare i32 @llvm.amdgcn.readfirstlane(i32) #0
declare i64 @llvm.amdgcn.icmp.i32(i32, i32, i32) #1
declare i64 @llvm.amdgcn.fcmp.i32(float, float, i32) #1
attributes #0 = { nounwind readnone }
attributes #1 = { nounwind readnone convergent }