blob: 6654f3322f0decd7e3f7d25ed6bca84fc8a12948 (
plain) (
blame)
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
|
//===----------------------------------------------------------------------===//
//
// Part of libcu++, the C++ Standard Library for your entire system,
// under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
// SPDX-FileCopyrightText: Copyright (c) 2025 NVIDIA CORPORATION & AFFILIATES.
//
//===----------------------------------------------------------------------===//
#ifndef _LIBCUDACXX___CMATH_ABS_H
#define _LIBCUDACXX___CMATH_ABS_H
#include <cuda/std/detail/__config>
#if defined(_CCCL_IMPLICIT_SYSTEM_HEADER_GCC)
# pragma GCC system_header
#elif defined(_CCCL_IMPLICIT_SYSTEM_HEADER_CLANG)
# pragma clang system_header
#elif defined(_CCCL_IMPLICIT_SYSTEM_HEADER_MSVC)
# pragma system_header
#endif // no system header
#include <cuda/std/__cmath/signbit.h>
#include <cuda/std/__concepts/concept_macros.h>
#include <cuda/std/__floating_point/fp.h>
#include <cuda/std/__type_traits/is_constant_evaluated.h>
#include <cuda/std/__type_traits/is_integral.h>
#include <cuda/std/cstdint>
// MSVC and clang cuda need the host side functions included
#if _CCCL_COMPILER(MSVC) || _CCCL_CUDA_COMPILER(CLANG)
# include <math.h>
#endif // _CCCL_COMPILER(MSVC) || _CCCL_CUDA_COMPILER(CLANG)
#include <cuda/std/__cccl/prologue.h>
_LIBCUDACXX_BEGIN_NAMESPACE_STD
// fabs
#if _CCCL_CHECK_BUILTIN(builtin_fabs) || _CCCL_COMPILER(GCC)
# define _CCCL_BUILTIN_FABSF(...) __builtin_fabsf(__VA_ARGS__)
# define _CCCL_BUILTIN_FABS(...) __builtin_fabs(__VA_ARGS__)
# define _CCCL_BUILTIN_FABSL(...) __builtin_fabsl(__VA_ARGS__)
#endif // _CCCL_CHECK_BUILTIN(builtin_fabs)
[[nodiscard]] _CCCL_API inline float fabs(float __x) noexcept
{
#if defined(_CCCL_BUILTIN_FABSF)
return _CCCL_BUILTIN_FABSF(__x);
#else // ^^^ _CCCL_BUILTIN_FABSF ^^^ / vvv !_CCCL_BUILTIN_FABSF vvv
return ::fabsf(__x);
#endif // ^^^ !_CCCL_BUILTIN_FABSF ^^^
}
[[nodiscard]] _CCCL_API inline float fabsf(float __x) noexcept
{
#if defined(_CCCL_BUILTIN_FABSF)
return _CCCL_BUILTIN_FABSF(__x);
#else // ^^^ _CCCL_BUILTIN_FABSF ^^^ / vvv !_CCCL_BUILTIN_FABSF vvv
return ::fabsf(__x);
#endif // ^^^ !_CCCL_BUILTIN_FABSF ^^^
}
[[nodiscard]] _CCCL_API inline double fabs(double __x) noexcept
{
#if defined(_CCCL_BUILTIN_FABS)
return _CCCL_BUILTIN_FABS(__x);
#else // ^^^ _CCCL_BUILTIN_FABS ^^^ / vvv !_CCCL_BUILTIN_FABS vvv
return ::fabs(__x);
#endif // ^^^ !_CCCL_BUILTIN_FABS ^^^
}
#if _CCCL_HAS_LONG_DOUBLE()
[[nodiscard]] _CCCL_API inline long double fabs(long double __x) noexcept
{
# if defined(_CCCL_BUILTIN_FABSL)
return _CCCL_BUILTIN_FABSL(__x);
# else // ^^^ _CCCL_BUILTIN_FABSL ^^^ / vvv !_CCCL_BUILTIN_FABSL vvv
return ::fabsl(__x);
# endif // ^^^ !_CCCL_BUILTIN_FABSL ^^^
}
[[nodiscard]] _CCCL_API inline long double fabsl(long double __x) noexcept
{
# if defined(_CCCL_BUILTIN_FABSL)
return _CCCL_BUILTIN_FABSL(__x);
# else // ^^^ _CCCL_BUILTIN_FABSL ^^^ / vvv !_CCCL_BUILTIN_FABSL vvv
return ::fabsl(__x);
# endif // ^^^ !_CCCL_BUILTIN_FABSL ^^^
}
#endif // _CCCL_HAS_LONG_DOUBLE()
template <class _Tp>
[[nodiscard]] _CCCL_API constexpr _Tp __fabs_impl(_Tp __x) noexcept
{
const auto __val = _CUDA_VSTD::__fp_get_storage(__x) & __fp_exp_mant_mask_of_v<_Tp>;
return _CUDA_VSTD::__fp_from_storage<_Tp>(static_cast<__fp_storage_of_t<_Tp>>(__val));
}
#if _CCCL_HAS_NVFP16()
[[nodiscard]] _CCCL_API constexpr __half fabs(__half __x) noexcept
{
// We cannot use `abs.f16` because it is not IEEE 754 compliant, see docs
return _CUDA_VSTD::__fabs_impl(__x);
}
#endif // _CCCL_HAS_NVFP16()
#if _CCCL_HAS_NVBF16()
[[nodiscard]] _CCCL_API constexpr __nv_bfloat16 fabs(__nv_bfloat16 __x) noexcept
{
// We cannot use `abs.bf16` because it is not IEEE 754 compliant, see docs
return _CUDA_VSTD::__fabs_impl(__x);
}
#endif // _CCCL_HAS_NVBF16()
#if _CCCL_HAS_NVFP8_E4M3()
[[nodiscard]] _CCCL_API constexpr __nv_fp8_e4m3 fabs(__nv_fp8_e4m3 __x) noexcept
{
return _CUDA_VSTD::__fabs_impl(__x);
}
#endif // _CCCL_HAS_NVFP8_E4M#()
#if _CCCL_HAS_NVFP8_E5M2()
[[nodiscard]] _CCCL_API constexpr __nv_fp8_e5m2 fabs(__nv_fp8_e5m2 __x) noexcept
{
return _CUDA_VSTD::__fabs_impl(__x);
}
#endif // _CCCL_HAS_NVFP8_E5M2()
#if _CCCL_HAS_NVFP8_E8M0()
[[nodiscard]] _CCCL_API constexpr __nv_fp8_e8m0 fabs(__nv_fp8_e8m0 __x) noexcept
{
return __x;
}
#endif // _CCCL_HAS_NVFP8_E8M0()
#if _CCCL_HAS_NVFP6_E2M3()
[[nodiscard]] _CCCL_API constexpr __nv_fp6_e2m3 fabs(__nv_fp6_e2m3 __x) noexcept
{
return _CUDA_VSTD::__fabs_impl(__x);
}
#endif // _CCCL_HAS_NVFP6_E2M3()
#if _CCCL_HAS_NVFP6_E3M2()
[[nodiscard]] _CCCL_API constexpr __nv_fp6_e3m2 fabs(__nv_fp6_e3m2 __x) noexcept
{
return _CUDA_VSTD::__fabs_impl(__x);
}
#endif // _CCCL_HAS_NVFP6_E3M2()
#if _CCCL_HAS_NVFP4_E2M1()
[[nodiscard]] _CCCL_API constexpr __nv_fp4_e2m1 fabs(__nv_fp4_e2m1 __x) noexcept
{
return _CUDA_VSTD::__fabs_impl(__x);
}
#endif // _CCCL_HAS_NVFP4_E2M1()
_CCCL_TEMPLATE(class _Tp)
_CCCL_REQUIRES(_CCCL_TRAIT(is_integral, _Tp))
[[nodiscard]] _CCCL_API inline double fabs(_Tp __val) noexcept
{
return _CUDA_VSTD::fabs(static_cast<double>(__val));
}
// abs
[[nodiscard]] _CCCL_API inline float abs(float __val) noexcept
{
return _CUDA_VSTD::fabsf(__val);
}
[[nodiscard]] _CCCL_API inline double abs(double __val) noexcept
{
return _CUDA_VSTD::fabs(__val);
}
#if _CCCL_HAS_LONG_DOUBLE()
[[nodiscard]] _CCCL_API inline long double abs(long double __val) noexcept
{
return _CUDA_VSTD::fabsl(__val);
}
#endif // _CCCL_HAS_LONG_DOUBLE()
#if _LIBCUDACXX_HAS_NVFP16()
[[nodiscard]] _CCCL_API constexpr __half abs(__half __x) noexcept
{
return _CUDA_VSTD::fabs(__x);
}
#endif // _LIBCUDACXX_HAS_NVFP16()
#if _LIBCUDACXX_HAS_NVBF16()
[[nodiscard]] _CCCL_API constexpr __nv_bfloat16 abs(__nv_bfloat16 __x) noexcept
{
return _CUDA_VSTD::fabs(__x);
}
#endif // _LIBCUDACXX_HAS_NVBF16()
#if _CCCL_HAS_NVFP8_E4M3()
[[nodiscard]] _CCCL_API constexpr __nv_fp8_e4m3 abs(__nv_fp8_e4m3 __x) noexcept
{
return _CUDA_VSTD::fabs(__x);
}
#endif // _CCCL_HAS_NVFP8_E4M3()
#if _CCCL_HAS_NVFP8_E5M2()
[[nodiscard]] _CCCL_API constexpr __nv_fp8_e5m2 abs(__nv_fp8_e5m2 __x) noexcept
{
return _CUDA_VSTD::fabs(__x);
}
#endif // _CCCL_HAS_NVFP8_E5M2()
#if _CCCL_HAS_NVFP8_E8M0()
[[nodiscard]] _CCCL_API constexpr __nv_fp8_e8m0 abs(__nv_fp8_e8m0 __x) noexcept
{
return _CUDA_VSTD::fabs(__x);
}
#endif // _CCCL_HAS_NVFP8_E8M0()
#if _CCCL_HAS_NVFP6_E2M3()
[[nodiscard]] _CCCL_API constexpr __nv_fp6_e2m3 abs(__nv_fp6_e2m3 __x) noexcept
{
return _CUDA_VSTD::fabs(__x);
}
#endif // _CCCL_HAS_NVFP6_E2M3()
#if _CCCL_HAS_NVFP6_E3M2()
[[nodiscard]] _CCCL_API constexpr __nv_fp6_e3m2 abs(__nv_fp6_e3m2 __x) noexcept
{
return _CUDA_VSTD::fabs(__x);
}
#endif // _CCCL_HAS_NVFP6_E3M2()
#if _CCCL_HAS_NVFP4_E2M1()
[[nodiscard]] _CCCL_API constexpr __nv_fp4_e2m1 abs(__nv_fp4_e2m1 __x) noexcept
{
return _CUDA_VSTD::fabs(__x);
}
#endif // _CCCL_HAS_NVFP4_E2M1()
_LIBCUDACXX_END_NAMESPACE_STD
#include <cuda/std/__cccl/epilogue.h>
#endif // _LIBCUDACXX___CMATH_ABS_H
|
