Skip to content

Commit 33315ef

Browse files
arsenmarsenm
committed
clang/AMDGPU: Don't set implicit arg attribute to default size
Since 2959e08, we conservatively assume all inputs are enabled by default. This isn't the best interface for controlling these anyway, since it's not granular and only allows trimming the last fields.
1 parent bed947f commit 33315ef

File tree

4 files changed

+29
-45
lines changed

4 files changed

+29
-45
lines changed

‎clang/lib/CodeGen/TargetInfo.cpp

Lines changed: 0 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -9304,16 +9304,9 @@ void AMDGPUTargetCodeGenInfo::setTargetAttributes(
93049304
if (FD)
93059305
setFunctionDeclAttributes(FD, F, M);
93069306

9307-
const bool IsOpenCLKernel =
9308-
M.getLangOpts().OpenCL && FD && FD->hasAttr<OpenCLKernelAttr>();
93099307
const bool IsHIPKernel =
93109308
M.getLangOpts().HIP && FD && FD->hasAttr<CUDAGlobalAttr>();
93119309

9312-
const bool IsOpenMP = M.getLangOpts().OpenMP && !FD;
9313-
if ((IsOpenCLKernel || IsHIPKernel || IsOpenMP) &&
9314-
(M.getTriple().getOS() == llvm::Triple::AMDHSA))
9315-
F->addFnAttr("amdgpu-implicitarg-num-bytes", "56");
9316-
93179310
if (IsHIPKernel)
93189311
F->addFnAttr("uniform-work-group-size", "true");
93199312

‎clang/test/CodeGenCUDA/amdgpu-hip-implicit-kernarg.cu

Lines changed: 0 additions & 8 deletions
This file was deleted.

‎clang/test/CodeGenOpenCL/amdgpu-attrs.cl

Lines changed: 25 additions & 26 deletions
Original file line numberDiff line numberDiff line change
@@ -1,5 +1,5 @@
11
// RUN: %clang_cc1 -triple amdgcn-amd-amdhsa -target-cpu tahiti -O0 -emit-llvm -o - %s | FileCheck %s
2-
// RUN: %clang_cc1 -triple amdgcn-- -target-cpu tahiti -O0 -emit-llvm -o - %s | FileCheck %s -check-prefix=NONAMDHSA
2+
// RUN: %clang_cc1 -triple amdgcn-- -target-cpu tahiti -O0 -emit-llvm -o - %s | FileCheck %s
33
// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -O0 -emit-llvm -verify -o - %s | FileCheck -check-prefix=X86 %s
44

55
__attribute__((amdgpu_flat_work_group_size(0, 0))) // expected-no-diagnostics
@@ -153,42 +153,41 @@ kernel void default_kernel() {
153153
// X86-NOT: "amdgpu-waves-per-eu"
154154
// X86-NOT: "amdgpu-num-vgpr"
155155
// X86-NOT: "amdgpu-num-sgpr"
156-
// X86-NOT: "amdgpu-implicitarg-num-bytes"
157-
// NONAMDHSA-NOT: "amdgpu-implicitarg-num-bytes"
156+
// CHECK-NOT: "amdgpu-implicitarg-num-bytes"
158157

159158
// CHECK-NOT: "amdgpu-flat-work-group-size"="0,0"
160159
// CHECK-NOT: "amdgpu-waves-per-eu"="0"
161160
// CHECK-NOT: "amdgpu-waves-per-eu"="0,0"
162161
// CHECK-NOT: "amdgpu-num-sgpr"="0"
163162
// CHECK-NOT: "amdgpu-num-vgpr"="0"
164163

165-
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-implicitarg-num-bytes"="56"
166-
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_64_64]] = {{.*}} "amdgpu-flat-work-group-size"="64,64" "amdgpu-implicitarg-num-bytes"="56"
167-
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_16_128]] = {{.*}} "amdgpu-flat-work-group-size"="16,128" "amdgpu-implicitarg-num-bytes"="56"
164+
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64]] = {{.*}} "amdgpu-flat-work-group-size"="32,64"
165+
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_64_64]] = {{.*}} "amdgpu-flat-work-group-size"="64,64"
166+
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_16_128]] = {{.*}} "amdgpu-flat-work-group-size"="16,128"
168167

169-
// CHECK-DAG: attributes [[WAVES_PER_EU_2]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-waves-per-eu"="2"
168+
// CHECK-DAG: attributes [[WAVES_PER_EU_2]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-waves-per-eu"="2"
170169

171-
// CHECK-DAG: attributes [[WAVES_PER_EU_2_4]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-waves-per-eu"="2,4"
172-
// CHECK-DAG: attributes [[NUM_SGPR_32]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-num-sgpr"="32"
173-
// CHECK-DAG: attributes [[NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-num-vgpr"="64"
170+
// CHECK-DAG: attributes [[WAVES_PER_EU_2_4]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-waves-per-eu"="2,4"
171+
// CHECK-DAG: attributes [[NUM_SGPR_32]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-num-sgpr"="32"
172+
// CHECK-DAG: attributes [[NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-num-vgpr"="64"
174173

175-
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_WAVES_PER_EU_2]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-waves-per-eu"="2"
176-
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_WAVES_PER_EU_2_4]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-waves-per-eu"="2,4"
177-
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_NUM_SGPR_32]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-num-sgpr"="32"
178-
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-num-vgpr"="64"
179-
// CHECK-DAG: attributes [[WAVES_PER_EU_2_NUM_SGPR_32]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-num-sgpr"="32" "amdgpu-waves-per-eu"="2"
180-
// CHECK-DAG: attributes [[WAVES_PER_EU_2_NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-num-vgpr"="64" "amdgpu-waves-per-eu"="2"
181-
// CHECK-DAG: attributes [[WAVES_PER_EU_2_4_NUM_SGPR_32]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-num-sgpr"="32" "amdgpu-waves-per-eu"="2,4"
182-
// CHECK-DAG: attributes [[WAVES_PER_EU_2_4_NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-num-vgpr"="64" "amdgpu-waves-per-eu"="2,4"
183-
// CHECK-DAG: attributes [[NUM_SGPR_32_NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-num-sgpr"="32" "amdgpu-num-vgpr"="64"
174+
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_WAVES_PER_EU_2]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-waves-per-eu"="2"
175+
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_WAVES_PER_EU_2_4]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-waves-per-eu"="2,4"
176+
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_NUM_SGPR_32]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-num-sgpr"="32"
177+
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-num-vgpr"="64"
178+
// CHECK-DAG: attributes [[WAVES_PER_EU_2_NUM_SGPR_32]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-num-sgpr"="32" "amdgpu-waves-per-eu"="2"
179+
// CHECK-DAG: attributes [[WAVES_PER_EU_2_NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-num-vgpr"="64" "amdgpu-waves-per-eu"="2"
180+
// CHECK-DAG: attributes [[WAVES_PER_EU_2_4_NUM_SGPR_32]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-num-sgpr"="32" "amdgpu-waves-per-eu"="2,4"
181+
// CHECK-DAG: attributes [[WAVES_PER_EU_2_4_NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-num-vgpr"="64" "amdgpu-waves-per-eu"="2,4"
182+
// CHECK-DAG: attributes [[NUM_SGPR_32_NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-num-sgpr"="32" "amdgpu-num-vgpr"="64"
184183

185-
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_WAVES_PER_EU_2_NUM_SGPR_32]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-num-sgpr"="32" "amdgpu-waves-per-eu"="2"
186-
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_WAVES_PER_EU_2_NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-num-vgpr"="64" "amdgpu-waves-per-eu"="2"
187-
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_WAVES_PER_EU_2_4_NUM_SGPR_32]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-num-sgpr"="32" "amdgpu-waves-per-eu"="2,4"
188-
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_WAVES_PER_EU_2_4_NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-num-vgpr"="64" "amdgpu-waves-per-eu"="2,4"
184+
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_WAVES_PER_EU_2_NUM_SGPR_32]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-num-sgpr"="32" "amdgpu-waves-per-eu"="2"
185+
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_WAVES_PER_EU_2_NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-num-vgpr"="64" "amdgpu-waves-per-eu"="2"
186+
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_WAVES_PER_EU_2_4_NUM_SGPR_32]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-num-sgpr"="32" "amdgpu-waves-per-eu"="2,4"
187+
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_WAVES_PER_EU_2_4_NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-num-vgpr"="64" "amdgpu-waves-per-eu"="2,4"
189188

190-
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_WAVES_PER_EU_2_NUM_SGPR_32_NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-num-sgpr"="32" "amdgpu-num-vgpr"="64" "amdgpu-waves-per-eu"="2"
191-
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_WAVES_PER_EU_2_4_NUM_SGPR_32_NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-implicitarg-num-bytes"="56" "amdgpu-num-sgpr"="32" "amdgpu-num-vgpr"="64" "amdgpu-waves-per-eu"="2,4"
189+
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_WAVES_PER_EU_2_NUM_SGPR_32_NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-num-sgpr"="32" "amdgpu-num-vgpr"="64" "amdgpu-waves-per-eu"="2"
190+
// CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64_WAVES_PER_EU_2_4_NUM_SGPR_32_NUM_VGPR_64]] = {{.*}} "amdgpu-flat-work-group-size"="32,64" "amdgpu-num-sgpr"="32" "amdgpu-num-vgpr"="64" "amdgpu-waves-per-eu"="2,4"
192191

193192
// CHECK-DAG: attributes [[A_FUNCTION]] = {{.*}}
194-
// CHECK-DAG: attributes [[DEFAULT_KERNEL_ATTRS]] = {{.*}} "amdgpu-flat-work-group-size"="1,256" "amdgpu-implicitarg-num-bytes"="56"
193+
// CHECK-DAG: attributes [[DEFAULT_KERNEL_ATTRS]] = {{.*}} "amdgpu-flat-work-group-size"="1,256"

‎clang/test/OpenMP/amdgcn-attributes.cpp

Lines changed: 4 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -32,10 +32,10 @@ int callable(int x) {
3232
return x + 1;
3333
}
3434

35-
// DEFAULT: attributes #0 = { convergent noinline norecurse nounwind optnone "amdgpu-implicitarg-num-bytes"="56" "frame-pointer"="none" "min-legal-vector-width"="0" "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
36-
// CPU: attributes #0 = { convergent noinline norecurse nounwind optnone "amdgpu-implicitarg-num-bytes"="56" "frame-pointer"="none" "min-legal-vector-width"="0" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="gfx900" "target-features"="+16-bit-insts,+ci-insts,+dpp,+flat-address-space,+gfx8-insts,+gfx9-insts,+s-memrealtime,+s-memtime-inst" }
37-
// NOIEEE: attributes #0 = { convergent noinline norecurse nounwind optnone "amdgpu-ieee"="false" "amdgpu-implicitarg-num-bytes"="56" "frame-pointer"="none" "min-legal-vector-width"="0" "no-nans-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
38-
// UNSAFEATOMIC: attributes #0 = { convergent noinline norecurse nounwind optnone "amdgpu-implicitarg-num-bytes"="56" "amdgpu-unsafe-fp-atomics"="true" "frame-pointer"="none" "min-legal-vector-width"="0" "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
35+
// DEFAULT: attributes #0 = { convergent noinline norecurse nounwind optnone "frame-pointer"="none" "min-legal-vector-width"="0" "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
36+
// CPU: attributes #0 = { convergent noinline norecurse nounwind optnone "frame-pointer"="none" "min-legal-vector-width"="0" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="gfx900" "target-features"="+16-bit-insts,+ci-insts,+dpp,+flat-address-space,+gfx8-insts,+gfx9-insts,+s-memrealtime,+s-memtime-inst" }
37+
// NOIEEE: attributes #0 = { convergent noinline norecurse nounwind optnone "amdgpu-ieee"="false" "frame-pointer"="none" "min-legal-vector-width"="0" "no-nans-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
38+
// UNSAFEATOMIC: attributes #0 = { convergent noinline norecurse nounwind optnone "amdgpu-unsafe-fp-atomics"="true" "frame-pointer"="none" "min-legal-vector-width"="0" "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
3939

4040
// DEFAULT: attributes #1 = { convergent mustprogress noinline nounwind optnone "frame-pointer"="none" "min-legal-vector-width"="0" "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
4141
// CPU: attributes #1 = { convergent mustprogress noinline nounwind optnone "frame-pointer"="none" "min-legal-vector-width"="0" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="gfx900" "target-features"="+16-bit-insts,+ci-insts,+dpp,+flat-address-space,+gfx8-insts,+gfx9-insts,+s-memrealtime,+s-memtime-inst" }

0 commit comments

Comments
 (0)