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
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
|
// This code contains NVIDIA Confidential Information and is disclosed to you
// under a form of NVIDIA software license agreement provided separately to you.
//
// Notice
// NVIDIA Corporation and its licensors retain all intellectual property and
// proprietary rights in and to this software and related documentation and
// any modifications thereto. Any use, reproduction, disclosure, or
// distribution of this software and related documentation without an express
// license agreement from NVIDIA Corporation is strictly prohibited.
//
// ALL NVIDIA DESIGN SPECIFICATIONS, CODE ARE PROVIDED "AS IS.". NVIDIA MAKES
// NO WARRANTIES, EXPRESSED, IMPLIED, STATUTORY, OR OTHERWISE WITH RESPECT TO
// THE MATERIALS, AND EXPRESSLY DISCLAIMS ALL IMPLIED WARRANTIES OF NONINFRINGEMENT,
// MERCHANTABILITY, AND FITNESS FOR A PARTICULAR PURPOSE.
//
// Information and code furnished is believed to be accurate and reliable.
// However, NVIDIA Corporation assumes no responsibility for the consequences of use of such
// information or for any infringement of patents or other rights of third parties that may
// result from its use. No license is granted by implication or otherwise under any patent
// or patent rights of NVIDIA Corporation. Details are subject to change without notice.
// This code supersedes and replaces all information previously supplied.
// NVIDIA Corporation products are not authorized for use as critical
// components in life support devices or systems without express written approval of
// NVIDIA Corporation.
//
// Copyright (c) 2008-2014 NVIDIA Corporation. All rights reserved.
// Copyright (c) 2004-2008 AGEIA Technologies, Inc. All rights reserved.
// Copyright (c) 2001-2004 NovodeX AG. All rights reserved.
#pragma once
// - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
// factory implementation
// - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
template <>
inline Simd4iFactory<const int&>::operator Simd4i() const
{
return _mm_set1_epi32(v);
}
inline Simd4iFactory<detail::FourTuple>::operator Simd4i() const
{
return reinterpret_cast<const Simd4i&>(v);
}
template <int i>
inline Simd4iFactory<detail::IntType<i> >::operator Simd4i() const
{
return _mm_set1_epi32(i);
}
template <>
inline Simd4iFactory<detail::IntType<0> >::operator Simd4i() const
{
return _mm_setzero_si128();
}
template <>
inline Simd4iFactory<const int*>::operator Simd4i() const
{
return _mm_loadu_si128(reinterpret_cast<const __m128i*>(v));
}
template <>
inline Simd4iFactory<detail::AlignedPointer<int> >::operator Simd4i() const
{
return _mm_load_si128(reinterpret_cast<const __m128i*>(v.ptr));
}
template <>
inline Simd4iFactory<detail::OffsetPointer<int> >::operator Simd4i() const
{
return _mm_load_si128(reinterpret_cast<const __m128i*>(reinterpret_cast<const char*>(v.ptr) + v.offset));
}
// - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
// expression template
// - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
template <>
inline ComplementExpr<Simd4i>::operator Simd4i() const
{
return _mm_andnot_si128(v, _mm_set1_epi32(0xffffffff));
}
Simd4i operator&(const ComplementExpr<Simd4i>& complement, const Simd4i& v)
{
return _mm_andnot_si128(complement.v, v);
}
Simd4i operator&(const Simd4i& v, const ComplementExpr<Simd4i>& complement)
{
return _mm_andnot_si128(complement.v, v);
}
// - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
// operator implementations
// - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Simd4i simdi::operator==(const Simd4i& v0, const Simd4i& v1)
{
return _mm_cmpeq_epi32(v0, v1);
}
Simd4i simdi::operator<(const Simd4i& v0, const Simd4i& v1)
{
return _mm_cmplt_epi32(v0, v1);
}
Simd4i simdi::operator>(const Simd4i& v0, const Simd4i& v1)
{
return _mm_cmpgt_epi32(v0, v1);
}
ComplementExpr<Simd4i> operator~(const Simd4i& v)
{
return ComplementExpr<Simd4i>(v);
}
Simd4i operator&(const Simd4i& v0, const Simd4i& v1)
{
return _mm_and_si128(v0, v1);
}
Simd4i operator|(const Simd4i& v0, const Simd4i& v1)
{
return _mm_or_si128(v0, v1);
}
Simd4i operator^(const Simd4i& v0, const Simd4i& v1)
{
return _mm_xor_si128(v0, v1);
}
Simd4i operator<<(const Simd4i& v, int shift)
{
return _mm_slli_epi32(v, shift);
}
Simd4i operator>>(const Simd4i& v, int shift)
{
return _mm_srli_epi32(v, shift);
}
Simd4i simdi::operator+(const Simd4i& v0, const Simd4i& v1)
{
return _mm_add_epi32(v0, v1);
}
Simd4i simdi::operator-(const Simd4i& v)
{
return _mm_sub_epi32(_mm_setzero_si128(), v);
}
Simd4i simdi::operator-(const Simd4i& v0, const Simd4i& v1)
{
return _mm_sub_epi32(v0, v1);
}
// - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
// function implementations
// - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Simd4i simd4i(const Simd4f& v)
{
return _mm_castps_si128(v);
}
Simd4i truncate(const Simd4f& v)
{
return _mm_cvttps_epi32(v);
}
int (&simdi::array(Simd4i& v))[4]
{
return reinterpret_cast<int(&)[4]>(v);
}
const int (&simdi::array(const Simd4i& v))[4]
{
return reinterpret_cast<const int(&)[4]>(v);
}
void store(int* ptr, const Simd4i& v)
{
_mm_storeu_si128(reinterpret_cast<__m128i*>(ptr), v);
}
void storeAligned(int* ptr, const Simd4i& v)
{
_mm_store_si128(reinterpret_cast<__m128i*>(ptr), v);
}
void storeAligned(int* ptr, unsigned int offset, const Simd4i& v)
{
_mm_store_si128(reinterpret_cast<__m128i*>(reinterpret_cast<char*>(ptr) + offset), v);
}
template <size_t i>
Simd4i splat(const Simd4i& v)
{
return _mm_shuffle_epi32(v, _MM_SHUFFLE(i, i, i, i));
}
Simd4i select(const Simd4i& mask, const Simd4i& v0, const Simd4i& v1)
{
return _mm_xor_si128(v1, _mm_and_si128(mask, _mm_xor_si128(v1, v0)));
}
int simdi::allEqual(const Simd4i& v0, const Simd4i& v1)
{
return allTrue(simdi::operator==(v0, v1));
}
int simdi::allEqual(const Simd4i& v0, const Simd4i& v1, Simd4i& outMask)
{
return allTrue(outMask = simdi::operator==(v0, v1));
}
int simdi::anyEqual(const Simd4i& v0, const Simd4i& v1)
{
return anyTrue(simdi::operator==(v0, v1));
}
int simdi::anyEqual(const Simd4i& v0, const Simd4i& v1, Simd4i& outMask)
{
return anyTrue(outMask = simdi::operator==(v0, v1));
}
int simdi::allGreater(const Simd4i& v0, const Simd4i& v1)
{
return allTrue(simdi::operator>(v0, v1));
}
int simdi::allGreater(const Simd4i& v0, const Simd4i& v1, Simd4i& outMask)
{
return allTrue(outMask = simdi::operator>(v0, v1));
}
int simdi::anyGreater(const Simd4i& v0, const Simd4i& v1)
{
return anyTrue(simdi::operator>(v0, v1));
}
int simdi::anyGreater(const Simd4i& v0, const Simd4i& v1, Simd4i& outMask)
{
return anyTrue(outMask = simdi::operator>(v0, v1));
}
int allTrue(const Simd4i& v)
{
return _mm_movemask_ps(_mm_castsi128_ps(v)) == 0xf;
}
int anyTrue(const Simd4i& v)
{
return _mm_movemask_ps(_mm_castsi128_ps(v));
}
|