blob: 7359fb0006991693d8e638ad9070490798e83c6f (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
|
// RUN: %clang_cc1 "-triple" "nvptx-nvidia-cuda" "-target-feature" "+ptx86" "-target-cpu" "sm_100a" -emit-llvm -fcuda-is-device -o - %s | FileCheck %s
// RUN: %clang_cc1 "-triple" "nvptx64-nvidia-cuda" "-target-feature" "+ptx86" "-target-cpu" "sm_100a" -emit-llvm -fcuda-is-device -o - %s | FileCheck %s
// CHECK: define{{.*}} void @_Z6kernelPf(ptr noundef %out_f)
__attribute__((global)) void kernel(float* out_f) {
float a = 3.0;
int i = 0;
out_f[i++] = __nvvm_redux_sync_fmin(a, 0xFF);
// CHECK: call contract float @llvm.nvvm.redux.sync.fmin
out_f[i++] = __nvvm_redux_sync_fmin_abs(a, 0xFF);
// CHECK: call contract float @llvm.nvvm.redux.sync.fmin.abs
out_f[i++] = __nvvm_redux_sync_fmin_NaN(a, 0xF0);
// CHECK: call contract float @llvm.nvvm.redux.sync.fmin.NaN
out_f[i++] = __nvvm_redux_sync_fmin_abs_NaN(a, 0x0F);
// CHECK: call contract float @llvm.nvvm.redux.sync.fmin.abs.NaN
out_f[i++] = __nvvm_redux_sync_fmax(a, 0xFF);
// CHECK: call contract float @llvm.nvvm.redux.sync.fmax
out_f[i++] = __nvvm_redux_sync_fmax_abs(a, 0x01);
// CHECK: call contract float @llvm.nvvm.redux.sync.fmax.abs
out_f[i++] = __nvvm_redux_sync_fmax_NaN(a, 0xF1);
// CHECK: call contract float @llvm.nvvm.redux.sync.fmax.NaN
out_f[i++] = __nvvm_redux_sync_fmax_abs_NaN(a, 0x10);
// CHECK: call contract float @llvm.nvvm.redux.sync.fmax.abs.NaN
// CHECK: ret void
}
|
