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
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
|
/* SPDX-License-Identifier: Apache-2.0
* Copyright 2011-2013 Intel Corporation
* Copyright 2011-2022 Blender Foundation */
#ifndef __UTIL_MATH_INT4_H__
#define __UTIL_MATH_INT4_H__
#ifndef __UTIL_MATH_H__
# error "Do not include this file directly, include util/types.h instead."
#endif
CCL_NAMESPACE_BEGIN
#ifndef __KERNEL_GPU__
ccl_device_inline int4 operator+(const int4 a, const int4 b)
{
# ifdef __KERNEL_SSE__
return int4(_mm_add_epi32(a.m128, b.m128));
# else
return make_int4(a.x + b.x, a.y + b.y, a.z + b.z, a.w + b.w);
# endif
}
ccl_device_inline int4 operator+=(int4 &a, const int4 b)
{
return a = a + b;
}
ccl_device_inline int4 operator-(const int4 a, const int4 b)
{
# ifdef __KERNEL_SSE__
return int4(_mm_sub_epi32(a.m128, b.m128));
# else
return make_int4(a.x - b.x, a.y - b.y, a.z - b.z, a.w - b.w);
# endif
}
ccl_device_inline int4 operator-=(int4 &a, const int4 b)
{
return a = a - b;
}
ccl_device_inline int4 operator>>(const int4 a, int i)
{
# ifdef __KERNEL_SSE__
return int4(_mm_srai_epi32(a.m128, i));
# else
return make_int4(a.x >> i, a.y >> i, a.z >> i, a.w >> i);
# endif
}
ccl_device_inline int4 operator<<(const int4 a, int i)
{
# ifdef __KERNEL_SSE__
return int4(_mm_slli_epi32(a.m128, i));
# else
return make_int4(a.x << i, a.y << i, a.z << i, a.w << i);
# endif
}
ccl_device_inline int4 operator<(const int4 a, const int4 b)
{
# ifdef __KERNEL_SSE__
return int4(_mm_cmplt_epi32(a.m128, b.m128));
# else
return make_int4(a.x < b.x, a.y < b.y, a.z < b.z, a.w < b.w);
# endif
}
ccl_device_inline int4 operator<(const int4 a, const int b)
{
return a < make_int4(b);
}
ccl_device_inline int4 operator==(const int4 a, const int4 b)
{
# ifdef __KERNEL_SSE__
return int4(_mm_cmpeq_epi32(a.m128, b.m128));
# else
return make_int4(a.x == b.x, a.y == b.y, a.z == b.z, a.w == b.w);
# endif
}
ccl_device_inline int4 operator==(const int4 a, const int b)
{
return a == make_int4(b);
}
ccl_device_inline int4 operator>=(const int4 a, const int4 b)
{
# ifdef __KERNEL_SSE__
return int4(_mm_xor_si128(_mm_set1_epi32(0xffffffff), _mm_cmplt_epi32(a.m128, b.m128)));
# else
return make_int4(a.x >= b.x, a.y >= b.y, a.z >= b.z, a.w >= b.w);
# endif
}
ccl_device_inline int4 operator>=(const int4 a, const int b)
{
return a >= make_int4(b);
}
ccl_device_inline int4 operator&(const int4 a, const int4 b)
{
# ifdef __KERNEL_SSE__
return int4(_mm_and_si128(a.m128, b.m128));
# else
return make_int4(a.x & b.x, a.y & b.y, a.z & b.z, a.w & b.w);
# endif
}
ccl_device_inline int4 operator|(const int4 a, const int4 b)
{
# ifdef __KERNEL_SSE__
return int4(_mm_or_si128(a.m128, b.m128));
# else
return make_int4(a.x | b.x, a.y | b.y, a.z | b.z, a.w | b.w);
# endif
}
ccl_device_inline int4 operator^(const int4 a, const int4 b)
{
# ifdef __KERNEL_SSE__
return int4(_mm_xor_si128(a.m128, b.m128));
# else
return make_int4(a.x ^ b.x, a.y ^ b.y, a.z ^ b.z, a.w ^ b.w);
# endif
}
ccl_device_inline int4 operator&(const int32_t a, const int4 b)
{
return make_int4(a) & b;
}
ccl_device_inline int4 operator&(const int4 a, const int32_t b)
{
return a & make_int4(b);
}
ccl_device_inline int4 operator|(const int32_t a, const int4 b)
{
return make_int4(a) | b;
}
ccl_device_inline int4 operator|(const int4 a, const int32_t b)
{
return a | make_int4(b);
}
ccl_device_inline int4 operator^(const int32_t a, const int4 b)
{
return make_int4(a) ^ b;
}
ccl_device_inline int4 operator^(const int4 a, const int32_t b)
{
return a ^ make_int4(b);
}
ccl_device_inline int4 &operator&=(int4 &a, const int4 b)
{
return a = a & b;
}
ccl_device_inline int4 &operator&=(int4 &a, const int32_t b)
{
return a = a & b;
}
ccl_device_inline int4 &operator|=(int4 &a, const int4 b)
{
return a = a | b;
}
ccl_device_inline int4 &operator|=(int4 &a, const int32_t b)
{
return a = a | b;
}
ccl_device_inline int4 &operator^=(int4 &a, const int4 b)
{
return a = a ^ b;
}
ccl_device_inline int4 &operator^=(int4 &a, const int32_t b)
{
return a = a ^ b;
}
ccl_device_inline int4 &operator<<=(int4 &a, const int32_t b)
{
return a = a << b;
}
ccl_device_inline int4 &operator>>=(int4 &a, const int32_t b)
{
return a = a >> b;
}
# ifdef __KERNEL_SSE__
ccl_device_forceinline const int4 srl(const int4 a, const int32_t b)
{
return int4(_mm_srli_epi32(a.m128, b));
}
# endif
ccl_device_inline int4 min(int4 a, int4 b)
{
# if defined(__KERNEL_SSE__) && defined(__KERNEL_SSE41__)
return int4(_mm_min_epi32(a.m128, b.m128));
# else
return make_int4(min(a.x, b.x), min(a.y, b.y), min(a.z, b.z), min(a.w, b.w));
# endif
}
ccl_device_inline int4 max(int4 a, int4 b)
{
# if defined(__KERNEL_SSE__) && defined(__KERNEL_SSE41__)
return int4(_mm_max_epi32(a.m128, b.m128));
# else
return make_int4(max(a.x, b.x), max(a.y, b.y), max(a.z, b.z), max(a.w, b.w));
# endif
}
ccl_device_inline int4 clamp(const int4 a, const int4 mn, const int4 mx)
{
return min(max(a, mn), mx);
}
ccl_device_inline int4 select(const int4 mask, const int4 a, const int4 b)
{
# ifdef __KERNEL_SSE__
return int4(_mm_or_si128(_mm_and_si128(mask, a), _mm_andnot_si128(mask, b)));
# else
return make_int4(
(mask.x) ? a.x : b.x, (mask.y) ? a.y : b.y, (mask.z) ? a.z : b.z, (mask.w) ? a.w : b.w);
# endif
}
ccl_device_inline int4 load_int4(const int *v)
{
# ifdef __KERNEL_SSE__
return int4(_mm_loadu_si128((__m128i *)v));
# else
return make_int4(v[0], v[1], v[2], v[3]);
# endif
}
#endif /* __KERNEL_GPU__ */
ccl_device_inline float4 cast(const int4 a)
{
#ifdef __KERNEL_SSE__
return float4(_mm_castsi128_ps(a));
#else
return make_float4(
__int_as_float(a.x), __int_as_float(a.y), __int_as_float(a.z), __int_as_float(a.w));
#endif
}
#ifdef __KERNEL_SSE__
ccl_device_forceinline int4 andnot(const int4 a, const int4 b)
{
return int4(_mm_andnot_si128(a.m128, b.m128));
}
template<size_t i0, size_t i1, size_t i2, size_t i3>
ccl_device_forceinline int4 shuffle(const int4 a)
{
# ifdef __KERNEL_NEON__
int32x4_t result = shuffle_neon<int32x4_t, i0, i1, i2, i3>(vreinterpretq_s32_m128i(a));
return int4(vreinterpretq_m128i_s32(result));
# else
return int4(_mm_shuffle_epi32(a, _MM_SHUFFLE(i3, i2, i1, i0)));
# endif
}
template<size_t i0, size_t i1, size_t i2, size_t i3>
ccl_device_forceinline int4 shuffle(const int4 a, const int4 b)
{
# ifdef __KERNEL_NEON__
int32x4_t result = shuffle_neon<int32x4_t, i0, i1, i2, i3>(vreinterpretq_s32_m128i(a),
vreinterpretq_s32_m128i(b));
return int4(vreinterpretq_m128i_s32(result));
# else
return int4(_mm_castps_si128(
_mm_shuffle_ps(_mm_castsi128_ps(a), _mm_castsi128_ps(b), _MM_SHUFFLE(i3, i2, i1, i0))));
# endif
}
template<size_t i0> ccl_device_forceinline int4 shuffle(const int4 b)
{
return shuffle<i0, i0, i0, i0>(b);
}
#endif
CCL_NAMESPACE_END
#endif /* __UTIL_MATH_INT4_H__ */
|
