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

[AMDGPU][GFX9] Set +fp32-denormals for >=gfx900 unless -cl-denorms-are-zero is set 

Differential Revision: https://reviews.llvm.org/D31482

llvm-svn: 300306
This commit is contained in:
Konstantin Zhuravlyov 2017-04-14 05:33:57 +00:00
parent c9a4fc0750
commit e668b1cd1e
2 changed files with 19 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
clang/lib/Basic/Targets.cpp
View File

@ -2112,9 +2112,12 @@ class AMDGPUTargetInfo final : public TargetInfo {
bool hasFP64:1; bool hasFP64:1;
bool hasFMAF:1; bool hasFMAF:1;
bool hasLDEXPF:1; bool hasLDEXPF:1;
bool hasFullSpeedFP32Denorms:1;
const AddrSpace AS; const AddrSpace AS;
static bool hasFullSpeedFMAF32(StringRef GPUName) {
return parseAMDGCNName(GPUName) >= GK_GFX9;
}
static bool isAMDGCN(const llvm::Triple &TT) { static bool isAMDGCN(const llvm::Triple &TT) {
return TT.getArch() == llvm::Triple::amdgcn; return TT.getArch() == llvm::Triple::amdgcn;
} }
@ -2130,7 +2133,6 @@ public:
hasFP64(false), hasFP64(false),
hasFMAF(false), hasFMAF(false),
hasLDEXPF(false), hasLDEXPF(false),
hasFullSpeedFP32Denorms(false),
AS(isGenericZero(Triple)){ AS(isGenericZero(Triple)){
if (getTriple().getArch() == llvm::Triple::amdgcn) { if (getTriple().getArch() == llvm::Triple::amdgcn) {
hasFP64 = true; hasFP64 = true;
@ -2200,7 +2202,8 @@ public:
hasFP64Denormals = true; hasFP64Denormals = true;
} }
if (!hasFP32Denormals) if (!hasFP32Denormals)
TargetOpts.Features.push_back((Twine(hasFullSpeedFP32Denorms && TargetOpts.Features.push_back(
(Twine(hasFullSpeedFMAF32(TargetOpts.CPU) &&
!CGOpts.FlushDenorm ? '+' : '-') + Twine("fp32-denormals")).str()); !CGOpts.FlushDenorm ? '+' : '-') + Twine("fp32-denormals")).str());
// Always do not flush fp64 or fp16 denorms. // Always do not flush fp64 or fp16 denorms.
if (!hasFP64Denormals && hasFP64) if (!hasFP64Denormals && hasFP64)

13
clang/test/CodeGenOpenCL/gfx9-fp32-denorms.cl Normal file
View File

@ -0,0 +1,13 @@
// REQUIRES: amdgpu-registered-target
// RUN: %clang_cc1 -triple amdgcn-unknown-unknown -target-cpu gfx900 -S -emit-llvm -o - %s | FileCheck --check-prefix=DEFAULT %s
// RUN: %clang_cc1 -triple amdgcn-unknown-unknown -target-cpu gfx900 -S -emit-llvm -o - -target-feature +fp32-denormals %s | FileCheck --check-prefix=FEATURE_FP32_DENORMALS_ON %s
// RUN: %clang_cc1 -triple amdgcn-unknown-unknown -target-cpu gfx900 -S -emit-llvm -o - -target-feature -fp32-denormals %s | FileCheck --check-prefix=FEATURE_FP32_DENORMALS_OFF %s
// RUN: %clang_cc1 -triple amdgcn-unknown-unknown -target-cpu gfx900 -S -emit-llvm -o - -cl-denorms-are-zero %s | FileCheck --check-prefix=OPT_DENORMS_ARE_ZERO %s
// DEFAULT: +fp32-denormals
// FEATURE_FP32_DENORMALS_ON: +fp32-denormals
// FEATURE_FP32_DENORMALS_OFF: -fp32-denormals
// OPT_DENORMS_ARE_ZERO: -fp32-denormals
kernel void gfx9_fp32_denorms() {}