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
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
|
//TEST(compute):SIMPLE(filecheck=CHECK_SPV): -target spirv
//TEST(compute):SIMPLE(filecheck=CHECK_GLSL): -target glsl
//TEST(compute):SIMPLE(filecheck=CHECK_METAL): -target metal
//TEST(compute):SIMPLE(filecheck=CHECK_WGSL): -target wgsl
//TEST(compute):COMPARE_COMPUTE_EX(filecheck-buffer=BUF):-vk -compute -shaderobj -emit-spirv-via-glsl -output-using-type
//TEST(compute):COMPARE_COMPUTE_EX(filecheck-buffer=BUF):-vk -compute -shaderobj -emit-spirv-directly -output-using-type -allow-glsl
//TEST(compute):COMPARE_COMPUTE_EX(filecheck-buffer=BUF):-slang -compute -profile cs_6_6 -dx12 -shaderobj -render-feature hardware-device -output-using-type
//TEST(compute):COMPARE_COMPUTE_EX(filecheck-buffer=BUF):-metal -compute -shaderobj -output-using-type
//TEST(compute):COMPARE_COMPUTE_EX(filecheck-buffer=BUF):-cpu -compute -shaderobj -output-using-type -allow-glsl
// 16 bit variants are not supported by WGSL.
//TEST(compute):COMPARE_COMPUTE_EX(filecheck-buffer=BUF):-wgpu -compute -shaderobj -xslang -DWGSL -render-feature half -output-using-type
// Debug info for inlining errors can be given out, so disable them for this test.
//TEST(compute):COMPARE_COMPUTE_EX(filecheck-buffer=BUF):-cuda -compute -g0 -output-using-type
//TEST_INPUT:ubuffer(data=[0 0 0 0 0 0 0 0 0 0], stride=4):out,name outputBuffer
RWStructuredBuffer<uint> outputBuffer;
bool verifyResult<T: IFloat> (T expected, T actual, T tolerance)
{
return (expected - tolerance) <= actual && actual <= (expected + tolerance);
}
bool verifyResultVector<T: IFloat, let N: int>(vector<T, N> expected, vector<T, N> actual, T tolerance = T(0.01))
{
bool isValid = true;
for (int i = 0; i < N; ++i)
isValid = isValid && verifyResult(expected[i], actual[i], tolerance);
return isValid;
}
[numthreads(1, 1, 1)]
void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)
{
uint index = 0U;
//
// Unorm4x8
//
// CHECK_SPV: OpExtInst{{.*}} PackUnorm4x8
// CHECK_SPV: OpExtInst{{.*}} UnpackUnorm4x8
//
// CHECK_GLSL: (unpackUnorm4x8
// CHECK_GLSL: (packUnorm4x8
//
// CHECK_METAL: unpack_unorm4x8_to_float
// CHECK_METAL: pack_float_to_unorm4x8
// CHECK_METAL: unpack_unorm4x8_to_half
// CHECK_METAL: pack_half_to_unorm4x8
//
// CHECK_WGSL: (unpack4x8unorm
// CHECK_WGSL: (pack4x8unorm
//
float4 unorm4x8Expected = float4(0.777, 0.233, 0.931, 0.777);
uint unorm4x8Packed = packUnorm4x8(unorm4x8Expected);
float4 unorm4x8Actual = unpackUnorm4x8ToFloat(unorm4x8Packed);
// BUF: 1
outputBuffer[index++] = verifyResultVector(unorm4x8Expected, unorm4x8Actual);
half4 unorm4x8HalfExpected = half4(0.123h, 0.456h, 0.789h, 0.321h);
uint unorm4x8HalfPacked = packUnorm4x8(unorm4x8HalfExpected);
half4 unorm4x8HalfActual = unpackUnorm4x8ToHalf(unorm4x8HalfPacked);
// BUF-NEXT: 1
outputBuffer[index++] = verifyResultVector(unorm4x8HalfExpected, unorm4x8HalfActual);
//
// Snorm4x8
//
// CHECK_SPV: OpExtInst{{.*}} PackSnorm4x8
// CHECK_SPV: OpExtInst{{.*}} UnpackSnorm4x8
//
// CHECK_GLSL: (unpackSnorm4x8
// CHECK_GLSL: (packSnorm4x8
//
// CHECK_METAL: unpack_snorm4x8_to_float
// CHECK_METAL: pack_float_to_snorm4x8
// CHECK_METAL: unpack_snorm4x8_to_half
// CHECK_METAL: pack_half_to_snorm4x8
//
// CHECK_WGSL: (unpack4x8snorm
// CHECK_WGSL: (pack4x8snorm
//
float4 snorm4x8Expected = float4(-0.500, 0.250, -0.750, 0.999);
uint snorm4x8Packed = packSnorm4x8(snorm4x8Expected);
float4 snorm4x8Actual = unpackSnorm4x8ToFloat(snorm4x8Packed);
// BUF-NEXT: 1
outputBuffer[index++] = verifyResultVector(snorm4x8Expected, snorm4x8Actual);
half4 snorm4x8HalfExpected = half4(-0.333h, 0.666h, -1.000h, 0.500h);
uint snorm4x8HalfPacked = packSnorm4x8(snorm4x8HalfExpected);
half4 snorm4x8HalfActual = unpackSnorm4x8ToHalf(snorm4x8HalfPacked);
// BUF-NEXT: 1
outputBuffer[index++] = verifyResultVector(snorm4x8HalfExpected, snorm4x8HalfActual);
//
// Unorm2x16
//
// CHECK_SPV: OpExtInst{{.*}} PackUnorm2x16
// CHECK_SPV: OpExtInst{{.*}} UnpackUnorm2x16
//
// CHECK_GLSL: (unpackUnorm2x16
// CHECK_GLSL: (packUnorm2x16
//
// CHECK_METAL: unpack_unorm2x16_to_float
// CHECK_METAL: pack_float_to_unorm2x16
// CHECK_METAL: unpack_unorm2x16_to_half
// CHECK_METAL: pack_half_to_unorm2x16
//
// CHECK_WGSL: (unpack2x16unorm
// CHECK_WGSL: (pack2x16unorm
//
float2 unorm2x16Expected = float2(0.1234, 0.8765);
uint unorm2x16Packed = packUnorm2x16(unorm2x16Expected);
float2 unorm2x16Actual = unpackUnorm2x16ToFloat(unorm2x16Packed);
// BUF-NEXT: 1
outputBuffer[index++] = verifyResultVector(unorm2x16Expected, unorm2x16Actual);
half2 unorm2x16HalfExpected = half2(0.7777h, 0.7777h);
uint unorm2x16HalfPacked = packUnorm2x16(unorm2x16HalfExpected);
half2 unorm2x16HalfActual = unpackUnorm2x16ToHalf(unorm2x16HalfPacked);
// BUF-NEXT: 1
outputBuffer[index++] = verifyResultVector(unorm2x16HalfExpected, unorm2x16HalfActual);
//
// Snorm2x16
//
// CHECK_SPV: OpExtInst{{.*}} UnpackSnorm2x16
// CHECK_SPV: OpExtInst{{.*}} PackSnorm2x16
//
// CHECK_GLSL: (unpackSnorm2x16
// CHECK_GLSL: (packSnorm2x16
//
// CHECK_METAL: unpack_snorm2x16_to_float
// CHECK_METAL: pack_float_to_snorm2x16
// CHECK_METAL: unpack_snorm2x16_to_half
// CHECK_METAL: pack_half_to_snorm2x16
//
// CHECK_WGSL: (unpack2x16snorm
// CHECK_WGSL: (pack2x16snorm
//
float2 snorm2x16Expected = float2(-0.4444, 0.8888);
uint snorm2x16Packed = packSnorm2x16(snorm2x16Expected);
float2 snorm2x16Actual = unpackSnorm2x16ToFloat(snorm2x16Packed);
// BUF-NEXT: 1
outputBuffer[index++] = verifyResultVector(snorm2x16Expected, snorm2x16Actual);
half2 snorm2x16HalfExpected = half2(-0.9999h, 0.3333h);
uint snorm2x16HalfPacked = packSnorm2x16(snorm2x16HalfExpected);
half2 snorm2x16HalfActual = unpackSnorm2x16ToHalf(snorm2x16HalfPacked);
// BUF-NEXT: 1
outputBuffer[index++] = verifyResultVector(snorm2x16HalfExpected, snorm2x16HalfActual);
//
// Half2x16
//
// CHECK_SPV: OpExtInst{{.*}} UnpackHalf2x16
// CHECK_SPV: OpExtInst{{.*}} PackHalf2x16
//
// CHECK_GLSL: (unpackHalf2x16
// CHECK_GLSL: (packHalf2x16
//
// CHECK_WGSL: (unpack2x16float
// CHECK_WGSL: (pack2x16float
//
float2 half2x16Expected = float2(130.32, -12.12);
uint half2x16Packed = packHalf2x16(half2x16Expected);
float2 half2x16Actual = unpackHalf2x16ToFloat(half2x16Packed);
// BUF-NEXT: 1
outputBuffer[index++] = verifyResultVector(half2x16Expected.x, half2x16Actual.x, 0.1);
half2 half2x16HalfExpected = half2(-2.1111h, 4450.9999h);
uint half2x16HalfPacked = packHalf2x16(half2x16HalfExpected);
half2 half2x16HalfActual = unpackHalf2x16ToHalf(half2x16HalfPacked);
// BUF-NEXT: 1
outputBuffer[index++] = verifyResultVector(half2x16HalfExpected, half2x16HalfActual);
}
|
