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
|
//TEST_CATEGORY(wave, compute)
//DISABLE_TEST:COMPARE_COMPUTE_EX:-cpu -compute -shaderobj
//DISABLE_TEST:COMPARE_COMPUTE_EX:-slang -compute -shaderobj
//TEST:COMPARE_COMPUTE_EX:-slang -compute -dx12 -profile cs_6_0 -shaderobj -render-feature hardware-device
//DISABLE_TEST(vulkan):COMPARE_COMPUTE_EX:-vk -compute -shaderobj
//TEST:COMPARE_COMPUTE_EX:-cuda -compute -capability cuda_sm_7_0 -shaderobj
//TEST_INPUT:ubuffer(data=[0 0 0 0 0 0 0 0], stride=4):out,name outputBuffer
RWStructuredBuffer<int> outputBuffer;
[numthreads(8, 1, 1)]
void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)
{
const int idx = int(dispatchThreadID.x);
// NOTE! dxc only supports bit ops on uint and associated types NOT int
// Also GLSL does not have built in support for int matrices. So we'll just try with float for now
// GLSL does not support matrix types for Wave like intrinsics
matrix<int, 2, 2> v0 = matrix<int, 2, 2>(idx + 1, idx + 2, idx + 3, idx + 4);
matrix<float, 2, 2> v1 = matrix<float, 2, 2>(v0) + matrix<float, 2, 2>(1, 1, 1, 1);
matrix<uint, 2, 2> uv0 = matrix<uint, 2, 2>(v0[0][0], v0[0][1], v0[1][0], v0[0][1]);
matrix<int, 2, 2> r0 = WaveActiveSum(v0);
matrix<float, 2, 2> r1 = WaveActiveSum(v1);
matrix<uint, 2, 2> r2 = WaveActiveBitXor(uv0);
matrix<uint, 2, 2> r3 = WaveActiveBitOr(uv0);
matrix<uint, 2, 2> r4 = WaveActiveBitAnd(uv0);
matrix<uint, 2, 2> r5 = r2 + r3 + r4;
matrix<int, 2, 2> r6 = matrix<int, 2, 2>(int(r5[0][0]), int(r5[0][1]), int(r5[1][0]), int(r5[1][1]));
matrix<int, 2, 2> r = r0 + matrix<int, 2, 2>(r1) + r6;
outputBuffer[idx] = r[0][0] + r[0][1] + r[1][0] + r[1][1];
}
|
