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
|
//TEST:COMPARE_COMPUTE(filecheck-buffer=CHECK):-cpu -output-using-type
//TEST:COMPARE_COMPUTE(filecheck-buffer=CHECK):-dx12 -output-using-type
//TEST:COMPARE_COMPUTE(filecheck-buffer=CHECK):-vk -output-using-type
//TEST:COMPARE_COMPUTE(filecheck-buffer=CHECK):-mtl -output-using-type
//TEST:COMPARE_COMPUTE(filecheck-buffer=CHECK):-cuda -output-using-type
//TEST:COMPARE_COMPUTE(filecheck-buffer=CHECK):-wgsl -output-using-type
// Test for dot product with 1-element vectors called from a generic function
// CHECK: 20
//TEST_INPUT:ubuffer(data=[0], stride=4):out,name=outputBuffer
RWStructuredBuffer<int> outputBuffer;
// Generic function that computes dot product for N-sized float vectors
__generic<let N : int>
float genericDotFloat(vector<float, N> a, vector<float, N> b)
{
return dot(a, b);
}
// Generic function that computes dot product for N-sized int vectors
__generic<let N : int>
int genericDotInt(vector<int, N> a, vector<int, N> b)
{
return dot(a, b);
}
// Generic function for testing with different N values
__generic<let N : int>
float testFloatDot(float value)
{
vector<float, N> vec1;
vector<float, N> vec2;
// Initialize all components to the same value
for (int i = 0; i < N; i++)
{
vec1[i] = value;
vec2[i] = value;
}
return genericDotFloat(vec1, vec2);
}
// Generic function for testing integer dot products
__generic<let N : int>
int testIntDot(int value)
{
vector<int, N> vec1;
vector<int, N> vec2;
// Initialize all components to the same value
for (int i = 0; i < N; i++)
{
vec1[i] = value;
vec2[i] = value;
}
return genericDotInt(vec1, vec2);
}
[numthreads(1, 1, 1)]
void computeMain()
{
// Test with N=1 (single element vectors) - this is the main test case
float floatResult1 = testFloatDot<1>(3.0); // 3.0 * 3.0 = 9.0
int intResult1 = testIntDot<1>(3); // 3 * 3 = 9
// Test with N=2 to ensure generic function works for other sizes
float floatResult2 = testFloatDot<2>(1.0); // (1.0*1.0 + 1.0*1.0) = 2.0
// Sum all results: 9 + 9 + 2 = 20
int result = int(floatResult1) + intResult1 + int(floatResult2);
outputBuffer[0] = result;
}
|
