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
|
/* SPDX-License-Identifier: Apache-2.0
* Copyright 2011-2022 Blender Foundation */
#ifndef __UTIL_TYPES_H__
#define __UTIL_TYPES_H__
#if !defined(__KERNEL_METAL__)
# include <stdlib.h>
#endif
/* Standard Integer Types */
#if !defined(__KERNEL_GPU__)
# include <stdint.h>
# include <stdio.h>
#endif
#include "util/defines.h"
#ifndef __KERNEL_GPU__
# include "util/optimization.h"
# include "util/simd.h"
#endif
CCL_NAMESPACE_BEGIN
/* Types
*
* Define simpler unsigned type names, and integer with defined number of bits.
* Also vector types, named to be compatible with OpenCL builtin types, while
* working for CUDA and C++ too. */
/* Shorter Unsigned Names */
typedef unsigned char uchar;
typedef unsigned int uint;
typedef unsigned short ushort;
/* Fixed Bits Types */
#ifndef __KERNEL_GPU__
/* Generic Memory Pointer */
typedef uint64_t device_ptr;
#endif /* __KERNEL_GPU__ */
ccl_device_inline size_t align_up(size_t offset, size_t alignment)
{
return (offset + alignment - 1) & ~(alignment - 1);
}
ccl_device_inline size_t divide_up(size_t x, size_t y)
{
return (x + y - 1) / y;
}
ccl_device_inline size_t round_up(size_t x, size_t multiple)
{
return ((x + multiple - 1) / multiple) * multiple;
}
ccl_device_inline size_t round_down(size_t x, size_t multiple)
{
return (x / multiple) * multiple;
}
ccl_device_inline bool is_power_of_two(size_t x)
{
return (x & (x - 1)) == 0;
}
CCL_NAMESPACE_END
/* Device side printf only tested on CUDA, may work on more GPU devices. */
#if !defined(__KERNEL_GPU__) || defined(__KERNEL_CUDA__)
# define __KERNEL_PRINTF__
#endif
ccl_device_inline void print_float(ccl_private const char *label, const float a)
{
#ifdef __KERNEL_PRINTF__
printf("%s: %.8f\n", label, (double)a);
#endif
}
/* Most GPU APIs matching native vector types, so we only need to implement them for
* CPU and oneAPI. */
#if defined(__KERNEL_GPU__) && !defined(__KERNEL_ONEAPI__)
# define __KERNEL_NATIVE_VECTOR_TYPES__
#endif
/* Vectorized types declaration. */
#include "util/types_uchar2.h"
#include "util/types_uchar3.h"
#include "util/types_uchar4.h"
#include "util/types_int2.h"
#include "util/types_int3.h"
#include "util/types_int4.h"
#include "util/types_int8.h"
#include "util/types_uint2.h"
#include "util/types_uint3.h"
#include "util/types_uint4.h"
#include "util/types_ushort4.h"
#include "util/types_float2.h"
#include "util/types_float3.h"
#include "util/types_float4.h"
#include "util/types_float8.h"
#include "util/types_spectrum.h"
/* Vectorized types implementation. */
#include "util/types_uchar2_impl.h"
#include "util/types_uchar3_impl.h"
#include "util/types_uchar4_impl.h"
#include "util/types_int2_impl.h"
#include "util/types_int3_impl.h"
#include "util/types_int4_impl.h"
#include "util/types_int8_impl.h"
#include "util/types_uint2_impl.h"
#include "util/types_uint3_impl.h"
#include "util/types_uint4_impl.h"
#include "util/types_float2_impl.h"
#include "util/types_float3_impl.h"
#include "util/types_float4_impl.h"
#include "util/types_float8_impl.h"
#endif /* __UTIL_TYPES_H__ */
|
