blob: 834b4c5cdf8a3a60b294094820cd7eccfdbe81a8 (
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
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
|
//TEST:SIMPLE(filecheck=CHECK_GLSL): -allow-glsl -stage compute -entry computeMain -target glsl
//TEST:SIMPLE(filecheck=CHECK_SPV): -allow-glsl -stage compute -entry computeMain -target spirv -emit-spirv-directly
//TEST:SIMPLE(filecheck=CHECK_HLSL): -allow-glsl -stage compute -entry computeMain -target hlsl -DTARGET_HLSL
// not testing cuda due to missing impl
//DISABLE_TEST:SIMPLE(filecheck=CHECK_CUDA): -allow-glsl -stage compute -entry computeMain -target cuda -DTARGET_CUDA
// not testing cpp due to missing impl
//DISABLE_TEST:SIMPLE(filecheck=CHECK_CPP): -allow-glsl -stage compute -entry computeMain -target cpp -DTARGET_CPP
//TEST(compute, vulkan):COMPARE_COMPUTE(filecheck-buffer=BUF):-vk -compute -entry computeMain -allow-glsl
//TEST(compute, vulkan):COMPARE_COMPUTE(filecheck-buffer=BUF):-vk -compute -entry computeMain -allow-glsl -emit-spirv-directly
//TEST(compute):COMPARE_COMPUTE(filecheck-buffer=BUF):-wgpu -compute -entry computeMain -allow-glsl -xslang -DWGPU
//TEST(compute):COMPARE_COMPUTE(filecheck-buffer=BUF):-metal -compute -entry computeMain -allow-glsl
#version 430
//TEST_INPUT:ubuffer(data=[0 0 0 0 0], stride=4):out,name=outputBuffer
buffer MyBlockName2
{
uint data[];
} outputBuffer;
layout(local_size_x = 32) in;
shared uint shareMem;
[[ForceInline]]
void _barrier()
{
#if !defined(WGPU)
subgroupBarrier();
#else
GroupMemoryBarrier();
#endif
}
[[ForceInline]]
void _memoryBarrier()
{
#if !defined(WGPU)
subgroupMemoryBarrier();
#else
GroupMemoryBarrier();
#endif
}
[[ForceInline]]
void _memoryBarrierShared()
{
#if !defined(WGPU)
subgroupMemoryBarrierShared();
#else
GroupMemoryBarrier();
#endif
}
[[ForceInline]]
void _memoryBarrierBuffer()
{
#if !defined(WGPU)
subgroupMemoryBarrierBuffer();
#else
GroupMemoryBarrier();
#endif
}
void computeMain()
{
// TODO: no test for image memory was done -- subgroupMemoryBarrierImage();
// tests are seperate since concurrency testing
shareMem = 100;
_memoryBarrierShared();
outputBuffer.data[0] = 1;
_barrier();
outputBuffer.data[0] = 2;
_barrier();
outputBuffer.data[1] = 1;
_memoryBarrier();
outputBuffer.data[1] = 2;
_barrier();
outputBuffer.data[2] = 1;
_memoryBarrierBuffer();
outputBuffer.data[2] = 2;
_barrier();
shareMem = 2;
_memoryBarrierShared();
outputBuffer.data[3] = shareMem;
_barrier();
if (subgroupElect()) {
outputBuffer.data[4] = gl_GlobalInvocationID.x + 2;
}
// CHECK_GLSL: void main(
// CHECK_SPV: OpEntryPoint
// CHECK_HLSL: void computeMain(
// CHECK_CUDA: void computeMain(
// CHECK_CPP: void _computeMain(
// BUF: 2
// BUF-NEXT: 2
// BUF-NEXT: 2
// BUF-NEXT: 2
// BUF-NEXT: 2
}
|
