[clang] 3a61245 - clang/AMDGPU: Assume denormals are enabled for the default target.

[Matt Arsenault via cfe-commits]

Author: Matt Arsenault
Date: 2020-04-15T09:17:26-04:00
New Revision: 3a612450508b314b7a6f4db142d0c619031d760e

URL: https://github.com/llvm/llvm-project/commit/3a612450508b314b7a6f4db142d0c619031d760e
DIFF: https://github.com/llvm/llvm-project/commit/3a612450508b314b7a6f4db142d0c619031d760e.diff

LOG: clang/AMDGPU: Assume denormals are enabled for the default target.

Since the default logic was based on having fast denormal/fma
features, and the default target has no features, we assumed flushing
by default. This fixes incorrectly assuming flushing in builds for
"generic" IR libraries.

The handling for no specified --cuda-gpu-arch in HIP is kind of
broken. Somewhere else forces a default target of gfx803, which does
not enable denormal handling by default. We don't see this default
switching here, so you'll end up with a different denormal mode
depending on whether you explicitly requested gfx803, or used it by
default.

Added: 
    

Modified: 
    clang/lib/Driver/ToolChains/AMDGPU.cpp
    clang/test/Driver/cl-denorms-are-zero.cl
    clang/test/Driver/cuda-flush-denormals-to-zero.cu

Removed: 
    


################################################################################
diff  --git a/clang/lib/Driver/ToolChains/AMDGPU.cpp b/clang/lib/Driver/ToolChains/AMDGPU.cpp
index 2a796f28403f..e8c0b824ace1 100644
--- a/clang/lib/Driver/ToolChains/AMDGPU.cpp
+++ b/clang/lib/Driver/ToolChains/AMDGPU.cpp
@@ -262,6 +262,11 @@ AMDGPUToolChain::TranslateArgs(const DerivedArgList &Args, StringRef BoundArch,
 
 bool AMDGPUToolChain::getDefaultDenormsAreZeroForTarget(
     llvm::AMDGPU::GPUKind Kind) {
+
+  // Assume nothing without a specific target.
+  if (Kind == llvm::AMDGPU::GK_NONE)
+    return false;
+
   const unsigned ArchAttr = llvm::AMDGPU::getArchAttrAMDGCN(Kind);
 
   // Default to enabling f32 denormals by default on subtargets where fma is

diff  --git a/clang/test/Driver/cl-denorms-are-zero.cl b/clang/test/Driver/cl-denorms-are-zero.cl
index 7774c0d60da8..e3fd095e5831 100644
--- a/clang/test/Driver/cl-denorms-are-zero.cl
+++ b/clang/test/Driver/cl-denorms-are-zero.cl
@@ -1,20 +1,24 @@
 // Slow FMAF and slow f32 denormals
-// RUN: %clang -### -target amdgcn--amdhsa -c -mcpu=pitcairn %s 2>&1 | FileCheck -check-prefixes=AMDGCN,AMDGCN-FLUSH %s
+// RUN: %clang -### -target amdgcn--amdhsa -nogpulib -c -mcpu=pitcairn %s 2>&1 | FileCheck -check-prefixes=AMDGCN,AMDGCN-FLUSH %s
 // RUN: %clang -### -cl-denorms-are-zero -o - -target amdgcn--amdhsa -c -mcpu=pitcairn %s 2>&1 | FileCheck -check-prefixes=AMDGCN,AMDGCN-FLUSH %s
 
 // Fast FMAF, but slow f32 denormals
-// RUN: %clang -### -target amdgcn--amdhsa -c -mcpu=tahiti %s 2>&1 | FileCheck -check-prefixes=AMDGCN,AMDGCN-FLUSH %s
+// RUN: %clang -### -target amdgcn--amdhsa -nogpulib -c -mcpu=tahiti %s 2>&1 | FileCheck -check-prefixes=AMDGCN,AMDGCN-FLUSH %s
 // RUN: %clang -### -cl-denorms-are-zero -o - -target amdgcn--amdhsa -c -mcpu=tahiti %s 2>&1 | FileCheck -check-prefixes=AMDGCN,AMDGCN-FLUSH %s
 
 // Fast F32 denormals, but slow FMAF
-// RUN: %clang -### -target amdgcn--amdhsa -c -mcpu=fiji %s 2>&1 | FileCheck -check-prefixes=AMDGCN,AMDGCN-FLUSH %s
+// RUN: %clang -### -target amdgcn--amdhsa -nogpulib -c -mcpu=fiji %s 2>&1 | FileCheck -check-prefixes=AMDGCN,AMDGCN-FLUSH %s
 // RUN: %clang -### -cl-denorms-are-zero -o - -target amdgcn--amdhsa -c -mcpu=fiji %s 2>&1 | FileCheck -check-prefixes=AMDGCN,AMDGCN-FLUSH %s
 
 // Fast F32 denormals and fast FMAF
-// RUN: %clang -### -target amdgcn--amdhsa -c -mcpu=gfx900 %s 2>&1 | FileCheck -check-prefixes=AMDGCN,AMDGCN-DENORM %s
-// RUN: %clang -### -cl-denorms-are-zero -o - -target amdgcn--amdhsa -c -mcpu=gfx900 %s 2>&1 | FileCheck -check-prefixes=AMDGCN,AMDGCN-FLUSH %s
+// RUN: %clang -### -target amdgcn--amdhsa -nogpulib -c -mcpu=gfx900 %s 2>&1 | FileCheck -check-prefixes=AMDGCN,AMDGCN-DENORM %s
+// RUN: %clang -### -cl-denorms-are-zero -o - -target amdgcn--amdhsa -nogpulib -c -mcpu=gfx900 %s 2>&1 | FileCheck -check-prefixes=AMDGCN,AMDGCN-FLUSH %s
+
+// Default target is artificial, but should assume a conservative default.
+// RUN: %clang -### -target amdgcn--amdhsa -nogpulib -c %s 2>&1 | FileCheck -check-prefixes=AMDGCN,AMDGCN-DENORM %s
+// RUN: %clang -### -cl-denorms-are-zero -o - -target amdgcn--amdhsa -nogpulib -c %s 2>&1 | FileCheck -check-prefixes=AMDGCN,AMDGCN-FLUSH %s
 
 // AMDGCN-FLUSH: "-fdenormal-fp-math-f32=preserve-sign,preserve-sign"
 
 // This should be omitted and default to ieee
-// AMDGCN-DENORM-NOT: "-fdenormal-fp-math-f32"
+// AMDGCN-DENORM-NOT: denormal-fp-math

diff  --git a/clang/test/Driver/cuda-flush-denormals-to-zero.cu b/clang/test/Driver/cuda-flush-denormals-to-zero.cu
index a515b5f8ca07..4a7b88823771 100644
--- a/clang/test/Driver/cuda-flush-denormals-to-zero.cu
+++ b/clang/test/Driver/cuda-flush-denormals-to-zero.cu
@@ -22,6 +22,8 @@
 // RUN: %clang -x hip -no-canonical-prefixes -### -target x86_64-linux-gnu -c -march=haswell --cuda-gpu-arch=gfx803 -nocudainc -nogpulib %s 2>&1 | FileCheck -check-prefix=FTZ %s
 // RUN: %clang -x hip -no-canonical-prefixes -### -target x86_64-linux-gnu -c -march=haswell --cuda-gpu-arch=gfx900 -nocudainc -nogpulib %s 2>&1 | FileCheck -check-prefix=NOFTZ %s
 
+// Test no subtarget, which should get the denormal setting of the default gfx803
+// RUN: %clang -x hip -no-canonical-prefixes -### -target x86_64-linux-gnu -c -march=haswell -nocudainc -nogpulib %s 2>&1 | FileCheck -check-prefix=FTZ %s
 
 // Test multiple offload archs with 
diff erent defaults.
 // RUN: %clang -x hip -no-canonical-prefixes -### -target x86_64-linux-gnu -c -march=haswell --cuda-gpu-arch=gfx803 --cuda-gpu-arch=gfx900 -nocudainc -nogpulib %s 2>&1 | FileCheck -check-prefix=MIXED-DEFAULT-MODE %s