yaxunl updated this revision to Diff 295326.
yaxunl added a comment.
simplifies wavefrontsize64 target feature
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D88370/new/
https://reviews.llvm.org/D88370
Files:
clang/lib/Basic/Targets/AMDGPU.cpp
clang/lib/Basic/Targets/AMDGPU.h
clang/lib/Driver/ToolChains/AMDGPU.cpp
clang/test/Driver/amdgpu-features.c
clang/test/Driver/amdgpu-macros.cl
clang/test/Driver/hip-macros.hip
clang/test/Driver/hip-toolchain-features.hip
Index: clang/test/Driver/hip-toolchain-features.hip
===================================================================
--- clang/test/Driver/hip-toolchain-features.hip
+++ clang/test/Driver/hip-toolchain-features.hip
@@ -37,8 +37,17 @@
// RUN: -mcumode -mcumode -mno-cumode -mwavefrontsize64 -mcumode \
// RUN: -mwavefrontsize64 -mno-wavefrontsize64 2>&1 \
// RUN: | FileCheck %s -check-prefix=DUP
-// DUP: {{.*}}clang{{.*}} "-target-feature" "-wavefrontsize16"
-// DUP-SAME: "-target-feature" "+wavefrontsize32"
-// DUP-SAME: "-target-feature" "-wavefrontsize64"
-// DUP-SAME: "-target-feature" "+cumode"
-// DUP: {{.*}}lld{{.*}} "-plugin-opt=-mattr=-wavefrontsize16,+wavefrontsize32,-wavefrontsize64,+cumode"
+// DUP: {{.*}}clang{{.*}} "-target-feature" "+cumode"
+// DUP-NOT: "-target-feature" "{{.*}}wavefrontsize16"
+// DUP-NOT: "-target-feature" "{{.*}}wavefrontsize32"
+// DUP-NOT: "-target-feature" "{{.*}}wavefrontsize64"
+// DUP: {{.*}}lld{{.*}} "-plugin-opt=-mattr=+cumode"
+
+// RUN: %clang -### -target x86_64-linux-gnu -fgpu-rdc -nogpulib \
+// RUN: --cuda-gpu-arch=gfx1010 %s \
+// RUN: -mno-wavefrontsize64 -mwavefrontsize64 2>&1 \
+// RUN: | FileCheck %s -check-prefix=WAVE64
+// WAVE64: {{.*}}clang{{.*}} "-target-feature" "+wavefrontsize64"
+// WAVE64-NOT: "-target-feature" "{{.*}}wavefrontsize16"
+// WAVE64-NOT: "-target-feature" "{{.*}}wavefrontsize32"
+// WAVE64: {{.*}}lld{{.*}} "-plugin-opt=-mattr=+wavefrontsize64"
Index: clang/test/Driver/hip-macros.hip
===================================================================
--- /dev/null
+++ clang/test/Driver/hip-macros.hip
@@ -0,0 +1,20 @@
+// RUN: %clang -E -dM --offload-arch=gfx906 -mwavefrontsize64 \
+// RUN: --cuda-device-only -nogpuinc -nogpulib \
+// RUN: %s 2>&1 | FileCheck --check-prefixes=WAVE64 %s
+// RUN: %clang -E -dM --offload-arch=gfx1010 -mwavefrontsize64 \
+// RUN: --cuda-device-only -nogpuinc -nogpulib \
+// RUN: %s 2>&1 | FileCheck --check-prefixes=WAVE64 %s
+// RUN: %clang -E -dM --offload-arch=gfx906 -mwavefrontsize64 \
+// RUN: --cuda-device-only -nogpuinc -nogpulib \
+// RUN: -mno-wavefrontsize64 %s 2>&1 | FileCheck --check-prefixes=WAVE64 %s
+// RUN: %clang -E -dM --offload-arch=gfx1010 -mwavefrontsize64 \
+// RUN: --cuda-device-only -nogpuinc -nogpulib \
+// RUN: -mno-wavefrontsize64 %s 2>&1 | FileCheck --check-prefixes=WAVE32 %s
+// RUN: %clang -E -dM --offload-arch=gfx906 -mno-wavefrontsize64 \
+// RUN: --cuda-device-only -nogpuinc -nogpulib \
+// RUN: -mwavefrontsize64 %s 2>&1 | FileCheck --check-prefixes=WAVE64 %s
+// RUN: %clang -E -dM --offload-arch=gfx1010 -mno-wavefrontsize64 \
+// RUN: --cuda-device-only -nogpuinc -nogpulib \
+// RUN: -mwavefrontsize64 %s 2>&1 | FileCheck --check-prefixes=WAVE64 %s
+// WAVE64-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// WAVE32-DAG: #define __AMDGCN_WAVEFRONT_SIZE 32
Index: clang/test/Driver/amdgpu-macros.cl
===================================================================
--- clang/test/Driver/amdgpu-macros.cl
+++ clang/test/Driver/amdgpu-macros.cl
@@ -346,4 +346,42 @@
// GFX1011-DAG: #define __amdgcn_processor__ "gfx1011"
// GFX1012-DAG: #define __amdgcn_processor__ "gfx1012"
// GFX1030-DAG: #define __amdgcn_processor__ "gfx1030"
-// GFX1031-DAG: #define __amdgcn_processor__ "gfx1031"
\ No newline at end of file
+// GFX1031-DAG: #define __amdgcn_processor__ "gfx1031"
+
+// GFX600-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// GFX601-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// GFX700-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// GFX701-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// GFX702-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// GFX703-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// GFX704-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// GFX801-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// GFX802-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// GFX803-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// GFX810-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// GFX900-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// GFX902-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// GFX904-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// GFX906-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// GFX908-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// GFX909-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// GFX1010-DAG: #define __AMDGCN_WAVEFRONT_SIZE 32
+// GFX1011-DAG: #define __AMDGCN_WAVEFRONT_SIZE 32
+// GFX1012-DAG: #define __AMDGCN_WAVEFRONT_SIZE 32
+// GFX1030-DAG: #define __AMDGCN_WAVEFRONT_SIZE 32
+// GFX1031-DAG: #define __AMDGCN_WAVEFRONT_SIZE 32
+
+// RUN: %clang -E -dM -target amdgcn -mcpu=gfx906 -mwavefrontsize64 \
+// RUN: %s 2>&1 | FileCheck --check-prefixes=WAVE64 %s
+// RUN: %clang -E -dM -target amdgcn -mcpu=gfx1010 -mwavefrontsize64 \
+// RUN: %s 2>&1 | FileCheck --check-prefixes=WAVE64 %s
+// RUN: %clang -E -dM -target amdgcn -mcpu=gfx906 -mwavefrontsize64 \
+// RUN: -mno-wavefrontsize64 %s 2>&1 | FileCheck --check-prefixes=WAVE64 %s
+// RUN: %clang -E -dM -target amdgcn -mcpu=gfx1010 -mwavefrontsize64 \
+// RUN: -mno-wavefrontsize64 %s 2>&1 | FileCheck --check-prefixes=WAVE32 %s
+// RUN: %clang -E -dM -target amdgcn -mcpu=gfx906 -mno-wavefrontsize64 \
+// RUN: -mwavefrontsize64 %s 2>&1 | FileCheck --check-prefixes=WAVE64 %s
+// RUN: %clang -E -dM -target amdgcn -mcpu=gfx1010 -mno-wavefrontsize64 \
+// RUN: -mwavefrontsize64 %s 2>&1 | FileCheck --check-prefixes=WAVE64 %s
+// WAVE64-DAG: #define __AMDGCN_WAVEFRONT_SIZE 64
+// WAVE32-DAG: #define __AMDGCN_WAVEFRONT_SIZE 32
Index: clang/test/Driver/amdgpu-features.c
===================================================================
--- clang/test/Driver/amdgpu-features.c
+++ clang/test/Driver/amdgpu-features.c
@@ -25,10 +25,16 @@
// NO-SRAM-ECC: "-target-feature" "-sram-ecc"
// RUN: %clang -### -target amdgcn-amdpal -mcpu=gfx1010 -mwavefrontsize64 %s 2>&1 | FileCheck --check-prefix=WAVE64 %s
-// WAVE64: "-target-feature" "-wavefrontsize16" "-target-feature" "-wavefrontsize32" "-target-feature" "+wavefrontsize64"
+// RUN: %clang -### -target amdgcn-amdpal -mcpu=gfx1010 -mno-wavefrontsize64 -mwavefrontsize64 %s 2>&1 | FileCheck --check-prefix=WAVE64 %s
+// WAVE64: "-target-feature" "+wavefrontsize64"
+// WAVE64-NOT: {{".*wavefrontsize16"}}
+// WAVE64-NOT: {{".*wavefrontsize32"}}
// RUN: %clang -### -target amdgcn -mcpu=gfx1010 -mno-wavefrontsize64 %s 2>&1 | FileCheck --check-prefix=NO-WAVE64 %s
-// NO-WAVE64: "-target-feature" "-wavefrontsize16" "-target-feature" "+wavefrontsize32" "-target-feature" "-wavefrontsize64"
+// RUN: %clang -### -target amdgcn -mcpu=gfx1010 -mwavefrontsize64 -mno-wavefrontsize64 %s 2>&1 | FileCheck --check-prefix=NO-WAVE64 %s
+// NO-WAVE64-NOT: {{".*wavefrontsize16"}}
+// NO-WAVE64-NOT: {{".*wavefrontsize32"}}
+// NO-WAVE64-NOT: {{".*wavefrontsize64"}}
// RUN: %clang -### -target amdgcn -mcpu=gfx1010 -mcumode %s 2>&1 | FileCheck --check-prefix=CUMODE %s
// CUMODE: "-target-feature" "+cumode"
Index: clang/lib/Driver/ToolChains/AMDGPU.cpp
===================================================================
--- clang/lib/Driver/ToolChains/AMDGPU.cpp
+++ clang/lib/Driver/ToolChains/AMDGPU.cpp
@@ -390,16 +390,9 @@
}
}
- if (Args.getLastArg(options::OPT_mwavefrontsize64)) {
- Features.push_back("-wavefrontsize16");
- Features.push_back("-wavefrontsize32");
+ if (Args.hasFlag(options::OPT_mwavefrontsize64,
+ options::OPT_mno_wavefrontsize64, false))
Features.push_back("+wavefrontsize64");
- }
- if (Args.getLastArg(options::OPT_mno_wavefrontsize64)) {
- Features.push_back("-wavefrontsize16");
- Features.push_back("+wavefrontsize32");
- Features.push_back("-wavefrontsize64");
- }
handleTargetFeaturesGroup(
Args, Features, options::OPT_m_amdgpu_Features_Group);
Index: clang/lib/Basic/Targets/AMDGPU.h
===================================================================
--- clang/lib/Basic/Targets/AMDGPU.h
+++ clang/lib/Basic/Targets/AMDGPU.h
@@ -41,6 +41,7 @@
llvm::AMDGPU::GPUKind GPUKind;
unsigned GPUFeatures;
+ unsigned WavefrontSize;
/// Target ID is device name followed by optional feature name postfixed
/// by plus or minus sign delimitted by colon, e.g. gfx908:xnack+:sram-ecc-.
@@ -407,6 +408,8 @@
getAllPossibleTargetIDFeatures(getTriple(), getArchNameAMDGCN(GPUKind));
llvm::for_each(Features, [&](const auto &F) {
assert(F.front() == '+' || F.front() == '-');
+ if (F == "+wavefrontsize64")
+ WavefrontSize = 64;
bool IsOn = F.front() == '+';
StringRef Name = StringRef(F).drop_front();
if (llvm::find(TargetIDFeatures, Name) == TargetIDFeatures.end())
Index: clang/lib/Basic/Targets/AMDGPU.cpp
===================================================================
--- clang/lib/Basic/Targets/AMDGPU.cpp
+++ clang/lib/Basic/Targets/AMDGPU.cpp
@@ -316,6 +316,7 @@
HasLegalHalfType = true;
HasFloat16 = true;
+ WavefrontSize = GPUFeatures & llvm::AMDGPU::FEATURE_WAVE32 ? 32 : 64;
// Set pointer width and alignment for target address space 0.
PointerWidth = PointerAlign = DataLayout->getPointerSizeInBits();
@@ -388,6 +389,8 @@
Builder.defineMacro("__HAS_FP64__");
if (hasFastFMA())
Builder.defineMacro("FP_FAST_FMA");
+
+ Builder.defineMacro("__AMDGCN_WAVEFRONT_SIZE", Twine(WavefrontSize));
}
void AMDGPUTargetInfo::setAuxTarget(const TargetInfo *Aux) {
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
