From e668b1cd1e78cb244b79d987aef725cc76c30e64 Mon Sep 17 00:00:00 2001
From: Konstantin Zhuravlyov <kzhuravl_dev@outlook.com>
Date: Fri, 14 Apr 2017 05:33:57 +0000
Subject: [AMDGPU][GFX9] Set +fp32-denormals for >=gfx900 unless
 -cl-denorms-are-zero is set

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

llvm-svn: 300306
---
 clang/lib/Basic/Targets.cpp | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

(limited to 'clang/lib/Basic/Targets.cpp')

diff --git a/clang/lib/Basic/Targets.cpp b/clang/lib/Basic/Targets.cpp
index fb22d402989..18f503d091c 100644
--- a/clang/lib/Basic/Targets.cpp
+++ b/clang/lib/Basic/Targets.cpp
@@ -2112,9 +2112,12 @@ class AMDGPUTargetInfo final : public TargetInfo {
   bool hasFP64:1;
   bool hasFMAF:1;
   bool hasLDEXPF:1;
-  bool hasFullSpeedFP32Denorms:1;
   const AddrSpace AS;
 
+  static bool hasFullSpeedFMAF32(StringRef GPUName) {
+    return parseAMDGCNName(GPUName) >= GK_GFX9;
+  }
+
   static bool isAMDGCN(const llvm::Triple &TT) {
     return TT.getArch() == llvm::Triple::amdgcn;
   }
@@ -2130,7 +2133,6 @@ public:
       hasFP64(false),
       hasFMAF(false),
       hasLDEXPF(false),
-      hasFullSpeedFP32Denorms(false),
       AS(isGenericZero(Triple)){
     if (getTriple().getArch() == llvm::Triple::amdgcn) {
       hasFP64 = true;
@@ -2200,7 +2202,8 @@ public:
         hasFP64Denormals = true;
     }
     if (!hasFP32Denormals)
-      TargetOpts.Features.push_back((Twine(hasFullSpeedFP32Denorms &&
+      TargetOpts.Features.push_back(
+          (Twine(hasFullSpeedFMAF32(TargetOpts.CPU) &&
           !CGOpts.FlushDenorm ? '+' : '-') + Twine("fp32-denormals")).str());
     // Always do not flush fp64 or fp16 denorms.
     if (!hasFP64Denormals && hasFP64)
-- 
cgit v1.2.3