1 // RUN: %clang_cc1 -triple amdgcn-amd-amdhsa \
2 // RUN: -fcuda-is-device -emit-llvm -o - -x hip %s \
3 // RUN: | FileCheck -check-prefixes=CHECK,DEFAULT %s
4 // RUN: %clang_cc1 -triple amdgcn-amd-amdhsa --gpu-max-threads-per-block=1024 \
5 // RUN: -fcuda-is-device -emit-llvm -o - -x hip %s \
6 // RUN: | FileCheck -check-prefixes=CHECK,MAX1024 %s
7 // RUN: %clang_cc1 -triple nvptx \
8 // RUN: -fcuda-is-device -emit-llvm -o - %s | FileCheck %s \
9 // RUN: -check-prefix=NAMD
10 // RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -emit-llvm \
11 // RUN: -verify -o - -x hip %s | FileCheck -check-prefix=NAMD %s
12
13 #include "Inputs/cuda.h"
14
flat_work_group_size_default()15 __global__ void flat_work_group_size_default() {
16 // CHECK: define amdgpu_kernel void @_Z28flat_work_group_size_defaultv() [[FLAT_WORK_GROUP_SIZE_DEFAULT:#[0-9]+]]
17 }
18
19 __attribute__((amdgpu_flat_work_group_size(32, 64))) // expected-no-diagnostics
flat_work_group_size_32_64()20 __global__ void flat_work_group_size_32_64() {
21 // CHECK: define amdgpu_kernel void @_Z26flat_work_group_size_32_64v() [[FLAT_WORK_GROUP_SIZE_32_64:#[0-9]+]]
22 }
23 __attribute__((amdgpu_waves_per_eu(2))) // expected-no-diagnostics
waves_per_eu_2()24 __global__ void waves_per_eu_2() {
25 // CHECK: define amdgpu_kernel void @_Z14waves_per_eu_2v() [[WAVES_PER_EU_2:#[0-9]+]]
26 }
27 __attribute__((amdgpu_num_sgpr(32))) // expected-no-diagnostics
num_sgpr_32()28 __global__ void num_sgpr_32() {
29 // CHECK: define amdgpu_kernel void @_Z11num_sgpr_32v() [[NUM_SGPR_32:#[0-9]+]]
30 }
31 __attribute__((amdgpu_num_vgpr(64))) // expected-no-diagnostics
num_vgpr_64()32 __global__ void num_vgpr_64() {
33 // CHECK: define amdgpu_kernel void @_Z11num_vgpr_64v() [[NUM_VGPR_64:#[0-9]+]]
34 }
35
36 // Make sure this is silently accepted on other targets.
37 // NAMD-NOT: "amdgpu-flat-work-group-size"
38 // NAMD-NOT: "amdgpu-waves-per-eu"
39 // NAMD-NOT: "amdgpu-num-vgpr"
40 // NAMD-NOT: "amdgpu-num-sgpr"
41
42 // DEFAULT-DAG: attributes [[FLAT_WORK_GROUP_SIZE_DEFAULT]] = {{.*}}"amdgpu-flat-work-group-size"="1,256"{{.*}}"uniform-work-group-size"="true"
43 // MAX1024-DAG: attributes [[FLAT_WORK_GROUP_SIZE_DEFAULT]] = {{.*}}"amdgpu-flat-work-group-size"="1,1024"
44 // CHECK-DAG: attributes [[FLAT_WORK_GROUP_SIZE_32_64]] = {{.*}}"amdgpu-flat-work-group-size"="32,64"
45 // CHECK-DAG: attributes [[WAVES_PER_EU_2]] = {{.*}}"amdgpu-waves-per-eu"="2"
46 // CHECK-DAG: attributes [[NUM_SGPR_32]] = {{.*}}"amdgpu-num-sgpr"="32"
47 // CHECK-DAG: attributes [[NUM_VGPR_64]] = {{.*}}"amdgpu-num-vgpr"="64"
48