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
|
// shader.slang
//TEST_INPUT:ubuffer(random(float, 4096, -1.0, 1.0), stride=4):name=ioBuffer
RWStructuredBuffer<float> convertBuffer(Ptr<gfx::BufferResource> x);
[shader("compute")]
[numthreads(4, 1, 1)]
void computeMain(uniform RWStructuredBuffer<float> ioBuffer, uint3 dispatchThreadID : SV_DispatchThreadID)
{
uint tid = dispatchThreadID.x;
float i = ioBuffer[tid];
float o = i < 0.5 ? (i + i) : sqrt(i);
ioBuffer[tid] = o;
}
// Forward declarations of gfx types
//
namespace gfx {
struct Device{};
struct BufferResource{};
struct ResourceView{};
struct TransientResourceHeap{};
struct PipelineState{};
struct ShaderProgram{};
}
// Forward declarations of cpp functions
//
Ptr<gfx::Device> createDevice();
Ptr<gfx::ShaderProgram> loadShaderProgram(Ptr<gfx::Device> device, String entryPoint, String module);
Ptr<gfx::BufferResource> createStructuredBuffer(
Ptr<gfx::Device> device,
float[4] initialData);
Ptr<gfx::ResourceView> createBufferView(
Ptr<gfx::Device> device,
Ptr<gfx::BufferResource> buffer);
Ptr<gfx::TransientResourceHeap> buildTransientHeap(
Ptr<gfx::Device> device);
Ptr<gfx::PipelineState> buildPipelineState(
Ptr<gfx::Device> device,
Ptr<gfx::ShaderProgram> shaderProgram);
void printInitialValues(float[4] initialArray, int length);
void dispatchComputation(
Ptr<gfx::Device> device,
Ptr<gfx::TransientResourceHeap> transientHeap,
Ptr<gfx::PipelineState> pipelineState,
Ptr<gfx::ResourceView> bufferView);
bool printOutputValues(
Ptr<gfx::Device> device,
Ptr<gfx::BufferResource> buffer,
int length);
public bool executeComputation() {
// We will hard-code the size of our initial array.
//
float initialArray[4] = { 3.0f, -20.0f, -6.0f, 8.0f };
// Declare functions
let device = createDevice();
let structuredBuffer = createStructuredBuffer(device, initialArray);
let bufferView = createBufferView(device, structuredBuffer);
__GPU_FOREACH(device, uint3(4, 1, 1), LAMBDA(uint3 dispatchThreadID)
{ computeMain(convertBuffer(structuredBuffer), dispatchThreadID) ; });
printInitialValues(initialArray, 4);
printOutputValues(device, structuredBuffer, 4);
return true;
}
|
