forked from apachecn/numpy-doc-zh
-
Notifications
You must be signed in to change notification settings - Fork 1
/
c-api.coremath.html
360 lines (357 loc) · 32 KB
/
c-api.coremath.html
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
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
<h1><span class="yiyi-st" id="yiyi-20">NumPy core libraries</span></h1>
<blockquote>
<p>原文：<a href="https://docs.scipy.org/doc/numpy/reference/c-api.coremath.html">https://docs.scipy.org/doc/numpy/reference/c-api.coremath.html</a></p>
<p>译者：<a href="https://github.com/wizardforcel">飞龙</a> <a href="http://usyiyi.cn/">UsyiyiCN</a></p>
<p>校对：（虚位以待）</p>
</blockquote>
<div class="versionadded">
<p><span class="yiyi-st" id="yiyi-21"><span class="versionmodified">版本1.3.0中的新功能。</span></span></p>
</div>
<p><span class="yiyi-st" id="yiyi-22">从numpy 1.3.0开始，我们正在努力将纯C，“计算”代码与python相关代码分离。</span><span class="yiyi-st" id="yiyi-23">目标是两个方面：使代码更清洁，并启用代码重用由numpy（scipy等）外的其他扩展。</span></p>
<div class="section" id="numpy-core-math-library">
<h2><span class="yiyi-st" id="yiyi-24">NumPy core math library</span></h2>
<p><span class="yiyi-st" id="yiyi-25">numpy核心数学库（'npymath'）是这个方向的第一步。</span><span class="yiyi-st" id="yiyi-26">这个库包含大多数与数学相关的C99功能，可以在没有很好支持C99的平台上使用。</span><span class="yiyi-st" id="yiyi-27">核心数学函数具有与C99相同的API，除了npy_ *前缀。</span></p>
<p><span class="yiyi-st" id="yiyi-28">The available functions are defined in <numpy npy_math.h=""> - please refer to this header when in doubt.</numpy></span></p>
<div class="section" id="floating-point-classification">
<h3><span class="yiyi-st" id="yiyi-29">Floating point classification</span></h3>
<dl class="var">
<dt id="c.NPY_NAN"><span class="yiyi-st" id="yiyi-30"> <code class="descname">NPY_NAN</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-31">这个宏定义为NaN（不是数字），并且保证未设置signbit（'positive'NaN）。</span><span class="yiyi-st" id="yiyi-32">相应的单精度宏和扩展精度宏可用后缀F和L.</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_INFINITY"><span class="yiyi-st" id="yiyi-33"> <code class="descname">NPY_INFINITY</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-34">此宏定义为正inf。</span><span class="yiyi-st" id="yiyi-35">相应的单精度宏和扩展精度宏可用后缀F和L.</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_PZERO"><span class="yiyi-st" id="yiyi-36"> <code class="descname">NPY_PZERO</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-37">此宏定义为正零。</span><span class="yiyi-st" id="yiyi-38">相应的单精度宏和扩展精度宏可用后缀F和L.</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_NZERO"><span class="yiyi-st" id="yiyi-39"> <code class="descname">NPY_NZERO</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-40">此宏定义为负零（即已设置符号位）。</span><span class="yiyi-st" id="yiyi-41">相应的单精度宏和扩展精度宏可用后缀F和L.</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_isnan"><span class="yiyi-st" id="yiyi-42"> int <code class="descname">npy_isnan</code><span class="sig-paren">(</span>x<span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-43">这是一个宏，并且等价于C99 isnan：适用于单精度，双精度和扩展精度，并返回非0值，x是NaN。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_isfinite"><span class="yiyi-st" id="yiyi-44"> int <code class="descname">npy_isfinite</code><span class="sig-paren">(</span>x<span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-45">这是一个宏，并且等效于C99 isfinite：适用于单精度，双精度和扩展精度，并返回非0值，x不是NaN也不是无穷大。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_isinf"><span class="yiyi-st" id="yiyi-46"> int <code class="descname">npy_isinf</code><span class="sig-paren">(</span>x<span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-47">这是一个宏，相当于C99 isinf：适用于单精度，双精度和扩展精度，返回一个非0值，x是无穷大（正和负）。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_signbit"><span class="yiyi-st" id="yiyi-48"> int <code class="descname">npy_signbit</code><span class="sig-paren">(</span>x<span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-49">这是一个宏，相当于C99 signbit：适用于单精度，双精度和扩展精度，并返回一个非0值，即x具有signbit集合（即数字为负）。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_copysign"><span class="yiyi-st" id="yiyi-50"> double <code class="descname">npy_copysign</code><span class="sig-paren">(</span>double<em> x</em>, double<em> y</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-51">这是一个等价于C99 copysign的函数：返回x与y相同的符号。</span><span class="yiyi-st" id="yiyi-52">适用于任何值，包括inf和nan。</span><span class="yiyi-st" id="yiyi-53">单和扩展精度可用后缀f和l。</span></p>
<div class="versionadded">
<p><span class="yiyi-st" id="yiyi-54"><span class="versionmodified">版本1.4.0中的新功能。</span></span></p>
</div>
</dd></dl>
</div>
<div class="section" id="useful-math-constants">
<h3><span class="yiyi-st" id="yiyi-55">Useful math constants</span></h3>
<p><span class="yiyi-st" id="yiyi-56">以下数学常数在npy_math.h中可用。</span><span class="yiyi-st" id="yiyi-57">还可以通过分别添加F和L后缀来获得单精度和扩展精度。</span></p>
<dl class="var">
<dt id="c.NPY_E"><span class="yiyi-st" id="yiyi-58"> <code class="descname">NPY_E</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-59">自然对数底（<img alt="e" class="math" src="../_images/math/ca907ab79fe0188f51d0a5db04642f8042f779d5.png" style="vertical-align: 0px">）</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_LOG2E"><span class="yiyi-st" id="yiyi-60"> <code class="descname">NPY_LOG2E</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-61">欧拉常数的基底2的对数（<img alt="\frac{\ln(e)}{\ln(2)}" class="math" src="../_images/math/f23b7585b585e4c1237a6424ef72595c19d3445a.png" style="vertical-align: -8px">）</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_LOG10E"><span class="yiyi-st" id="yiyi-62"> <code class="descname">NPY_LOG10E</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-63">欧拉常数（<img alt="\frac{\ln(e)}{\ln(10)}" class="math" src="../_images/math/257bb150bb440e2c2068e77118d1886097d1619a.png" style="vertical-align: -8px">）的以10为底的对数</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_LOGE2"><span class="yiyi-st" id="yiyi-64"> <code class="descname">NPY_LOGE2</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-65">2的自然对数（<img alt="\ln(2)" class="math" src="../_images/math/ee2815119b4f9fbcb9e964b6252313c8bd075b23.png" style="vertical-align: -4px">）</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_LOGE10"><span class="yiyi-st" id="yiyi-66"> <code class="descname">NPY_LOGE10</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-67">10的自然对数（<img alt="\ln(10)" class="math" src="../_images/math/03eb2554999615a1e4e44880b814ca24bfb3fda8.png" style="vertical-align: -4px">）</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_PI"><span class="yiyi-st" id="yiyi-68"> <code class="descname">NPY_PI</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-69">Pi（<img alt="\pi" class="math" src="../_images/math/7321e712e2b80bb4b00be0d9faaa525e3f53cde4.png" style="vertical-align: 0px">）</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_PI_2"><span class="yiyi-st" id="yiyi-70"> <code class="descname">NPY_PI_2</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-71">Pi除以2（<img alt="\frac{\pi}{2}" class="math" src="../_images/math/4204552f1f6533576c2b1ec914f4abf1cdaeddcd.png" style="vertical-align: -5px">）</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_PI_4"><span class="yiyi-st" id="yiyi-72"> <code class="descname">NPY_PI_4</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-73">Pi除以4（<img alt="\frac{\pi}{4}" class="math" src="../_images/math/1f4f2fabb9b30baa436bd54998b80a49f601aa17.png" style="vertical-align: -5px">）</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_1_PI"><span class="yiyi-st" id="yiyi-74"> <code class="descname">NPY_1_PI</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-75">p1（<img alt="\frac{1}{\pi}" class="math" src="../_images/math/4c062e7a23e76e4d2667da92fd0146485cabf899.png" style="vertical-align: -5px">）的倒数</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_2_PI"><span class="yiyi-st" id="yiyi-76"> <code class="descname">NPY_2_PI</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-77">两倍于pi的倒数（<img alt="\frac{2}{\pi}" class="math" src="../_images/math/862e4f6aea189d29854606d9c6596981cdebae3c.png" style="vertical-align: -5px">）</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_EULER"><span class="yiyi-st" id="yiyi-78"> <code class="descname">NPY_EULER</code></span></dt>
<dd><dl class="docutils">
<dt><span class="yiyi-st" id="yiyi-79">欧拉常数</span></dt>
<dd></dd>
</dl>
</dd></dl>
</div>
<div class="section" id="low-level-floating-point-manipulation">
<h3><span class="yiyi-st" id="yiyi-80">Low-level floating point manipulation</span></h3>
<p><span class="yiyi-st" id="yiyi-81">这些可用于精确的浮点比较。</span></p>
<dl class="function">
<dt id="c.npy_nextafter"><span class="yiyi-st" id="yiyi-82"> double <code class="descname">npy_nextafter</code><span class="sig-paren">(</span>double<em> x</em>, double<em> y</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-83">这是相当于C99 nextafter的函数：从y的方向返回下一个可表示的浮点值。</span><span class="yiyi-st" id="yiyi-84">单和扩展精度可用后缀f和l。</span></p>
<div class="versionadded">
<p><span class="yiyi-st" id="yiyi-85"><span class="versionmodified">版本1.4.0中的新功能。</span></span></p>
</div>
</dd></dl>
<dl class="function">
<dt id="c.npy_spacing"><span class="yiyi-st" id="yiyi-86"> double <code class="descname">npy_spacing</code><span class="sig-paren">(</span>double<em> x</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-87">这是一个等价于Fortran本质的函数。</span><span class="yiyi-st" id="yiyi-88">从x返回x和下一个可表示浮点值之间的距离，例如spacing（1）== eps。</span><span class="yiyi-st" id="yiyi-89">纳米的间距和+/- inf返回纳米。</span><span class="yiyi-st" id="yiyi-90">单和扩展精度可用后缀f和l。</span></p>
<div class="versionadded">
<p><span class="yiyi-st" id="yiyi-91"><span class="versionmodified">版本1.4.0中的新功能。</span></span></p>
</div>
</dd></dl>
<dl class="function">
<dt id="c.npy_set_floatstatus_divbyzero"><span class="yiyi-st" id="yiyi-92"> void <code class="descname">npy_set_floatstatus_divbyzero</code><span class="sig-paren">(</span><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-93">设置除零浮点异常</span></p>
<div class="versionadded">
<p><span class="yiyi-st" id="yiyi-94"><span class="versionmodified">版本1.6.0中的新功能。</span></span></p>
</div>
</dd></dl>
<dl class="function">
<dt id="c.npy_set_floatstatus_overflow"><span class="yiyi-st" id="yiyi-95"> void <code class="descname">npy_set_floatstatus_overflow</code><span class="sig-paren">(</span><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-96">设置溢出浮点异常</span></p>
<div class="versionadded">
<p><span class="yiyi-st" id="yiyi-97"><span class="versionmodified">版本1.6.0中的新功能。</span></span></p>
</div>
</dd></dl>
<dl class="function">
<dt id="c.npy_set_floatstatus_underflow"><span class="yiyi-st" id="yiyi-98"> void <code class="descname">npy_set_floatstatus_underflow</code><span class="sig-paren">(</span><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-99">设置下溢浮点异常</span></p>
<div class="versionadded">
<p><span class="yiyi-st" id="yiyi-100"><span class="versionmodified">版本1.6.0中的新功能。</span></span></p>
</div>
</dd></dl>
<dl class="function">
<dt id="c.npy_set_floatstatus_invalid"><span class="yiyi-st" id="yiyi-101"> void <code class="descname">npy_set_floatstatus_invalid</code><span class="sig-paren">(</span><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-102">设置无效浮点异常</span></p>
<div class="versionadded">
<p><span class="yiyi-st" id="yiyi-103"><span class="versionmodified">版本1.6.0中的新功能。</span></span></p>
</div>
</dd></dl>
<dl class="function">
<dt id="c.npy_get_floatstatus"><span class="yiyi-st" id="yiyi-104"> int <code class="descname">npy_get_floatstatus</code><span class="sig-paren">(</span><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-105">获取浮点状态。</span><span class="yiyi-st" id="yiyi-106">返回具有以下可能标志的位掩码：</span></p>
<ul class="simple">
<li><span class="yiyi-st" id="yiyi-107">NPY_FPE_DIVIDEBYZERO</span></li>
<li><span class="yiyi-st" id="yiyi-108">NPY_FPE_OVERFLOW</span></li>
<li><span class="yiyi-st" id="yiyi-109">NPY_FPE_UNDERFLOW</span></li>
<li><span class="yiyi-st" id="yiyi-110">NPY_FPE_INVALID</span></li>
</ul>
<div class="versionadded">
<p><span class="yiyi-st" id="yiyi-111"><span class="versionmodified">版本1.9.0中的新功能。</span></span></p>
</div>
</dd></dl>
<dl class="function">
<dt id="c.npy_clear_floatstatus"><span class="yiyi-st" id="yiyi-112"> int <code class="descname">npy_clear_floatstatus</code><span class="sig-paren">(</span><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-113">清除浮点状态。</span><span class="yiyi-st" id="yiyi-114">返回先前的状态掩码。</span></p>
<div class="versionadded">
<p><span class="yiyi-st" id="yiyi-115"><span class="versionmodified">版本1.9.0中的新功能。</span></span></p>
</div>
</dd></dl>
</div>
<div class="section" id="complex-functions">
<h3><span class="yiyi-st" id="yiyi-116">Complex functions</span></h3>
<div class="versionadded">
<p><span class="yiyi-st" id="yiyi-117"><span class="versionmodified">版本1.4.0中的新功能。</span></span></p>
</div>
<p><span class="yiyi-st" id="yiyi-118">已添加C99-样复合物功能。</span><span class="yiyi-st" id="yiyi-119">如果你想实现便携式C扩展，那么可以使用。</span><span class="yiyi-st" id="yiyi-120">由于我们仍然支持没有C99复杂类型的平台，你需要限制为C90兼容语法，例如：</span></p>
<div class="highlight-c"><div class="highlight"><pre><span></span><span class="cm">/* a = 1 + 2i \*/</span>
<span class="n">npy_complex</span> <span class="n">a</span> <span class="o">=</span> <span class="n">npy_cpack</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">);</span>
<span class="n">npy_complex</span> <span class="n">b</span><span class="p">;</span>
<span class="n">b</span> <span class="o">=</span> <span class="n">npy_log</span><span class="p">(</span><span class="n">a</span><span class="p">);</span>
</pre></div>
</div>
</div>
<div class="section" id="linking-against-the-core-math-library-in-an-extension">
<h3><span class="yiyi-st" id="yiyi-121">Linking against the core math library in an extension</span></h3>
<div class="versionadded">
<p><span class="yiyi-st" id="yiyi-122"><span class="versionmodified">版本1.4.0中的新功能。</span></span></p>
</div>
<p><span class="yiyi-st" id="yiyi-123">要在您自己的扩展中使用核心数学库，您需要在setup.py中的扩展中添加npymath编译和链接选项：</span></p>
<div class="highlight-default"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span> <span class="nn">numpy.distutils.misc_util</span> <span class="k">import</span> <span class="n">get_info</span>
<span class="gp">>>> </span><span class="n">info</span> <span class="o">=</span> <span class="n">get_info</span><span class="p">(</span><span class="s1">'npymath'</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">config</span><span class="o">.</span><span class="n">add_extension</span><span class="p">(</span><span class="s1">'foo'</span><span class="p">,</span> <span class="n">sources</span><span class="o">=</span><span class="p">[</span><span class="s1">'foo.c'</span><span class="p">],</span> <span class="n">extra_info</span><span class="o">=</span><span class="n">info</span><span class="p">)</span>
</pre></div>
</div>
<p><span class="yiyi-st" id="yiyi-124">换句话说，info的使用与使用blas_info和co时完全相同。</span></p>
</div>
<div class="section" id="half-precision-functions">
<h3><span class="yiyi-st" id="yiyi-125">Half-precision functions</span></h3>
<div class="versionadded">
<p><span class="yiyi-st" id="yiyi-126"><span class="versionmodified">版本2.0.0中的新功能。</span></span></p>
</div>
<p><span class="yiyi-st" id="yiyi-127">The header file <numpy halffloat.h=""> provides functions to work with IEEE 754-2008 16-bit floating point values. </numpy></span><span class="yiyi-st" id="yiyi-128">虽然此格式通常不用于数值计算，但它对于存储需要浮点但不需要很多精度的值很有用。</span><span class="yiyi-st" id="yiyi-129">它也可以用作一个教育工具来理解浮点舍入误差的性质。</span></p>
<p><span class="yiyi-st" id="yiyi-130">像其他类型一样，NumPy包括16位浮点型的typedef npy_half。</span><span class="yiyi-st" id="yiyi-131">与大多数其他类型不同，你不能使用它作为C中的正常类型，因为is是npy_uint16的typedef。</span><span class="yiyi-st" id="yiyi-132">例如，1.0看起来像0x3c00到C，如果你做不同的有符号零之间的等式比较，你会得到-0.0！= 0.0（0x8000！= 0x0000），这是不正确的。</span></p>
<p><span class="yiyi-st" id="yiyi-133">For these reasons, NumPy provides an API to work with npy_half values accessible by including <numpy halffloat.h=""> and linking to ‘npymath’. </numpy></span><span class="yiyi-st" id="yiyi-134">对于未直接提供的函数（如算术运算），首选方法是转换为float或double并再次返回，如以下示例所示。</span></p>
<div class="highlight-c"><div class="highlight"><pre><span></span><span class="n">npy_half</span> <span class="nf">sum</span><span class="p">(</span><span class="kt">int</span> <span class="n">n</span><span class="p">,</span> <span class="n">npy_half</span> <span class="o">*</span><span class="n">array</span><span class="p">)</span> <span class="p">{</span>
<span class="kt">float</span> <span class="n">ret</span> <span class="o">=</span> <span class="mi">0</span><span class="p">;</span>
<span class="k">while</span><span class="p">(</span><span class="n">n</span><span class="o">--</span><span class="p">)</span> <span class="p">{</span>
<span class="n">ret</span> <span class="o">+=</span> <span class="n">npy_half_to_float</span><span class="p">(</span><span class="o">*</span><span class="n">array</span><span class="o">++</span><span class="p">);</span>
<span class="p">}</span>
<span class="k">return</span> <span class="n">npy_float_to_half</span><span class="p">(</span><span class="n">ret</span><span class="p">);</span>
<span class="p">}</span>
</pre></div>
</div>
<p><span class="yiyi-st" id="yiyi-135">外部链接：</span></p>
<ul class="simple">
<li><span class="yiyi-st" id="yiyi-136"><a class="reference external" href="http://ieeexplore.ieee.org/servlet/opac?punumber=4610933">754-2008 IEEE浮点算术标准</a></span></li>
<li><span class="yiyi-st" id="yiyi-137"><a class="reference external" href="http://en.wikipedia.org/wiki/Half_precision_floating-point_format">半精度浮动维基百科文章</a>。</span></li>
<li><span class="yiyi-st" id="yiyi-138"><a class="reference external" href="http://www.opengl.org/registry/specs/ARB/half_float_pixel.txt">OpenGL半浮动像素支持</a></span></li>
<li><span class="yiyi-st" id="yiyi-139"><a class="reference external" href="http://www.openexr.com/about.html">OpenEXR图片格式</a>。</span></li>
</ul>
<dl class="var">
<dt id="c.NPY_HALF_ZERO"><span class="yiyi-st" id="yiyi-140"> <code class="descname">NPY_HALF_ZERO</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-141">此宏定义为正零。</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_HALF_PZERO"><span class="yiyi-st" id="yiyi-142"> <code class="descname">NPY_HALF_PZERO</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-143">此宏定义为正零。</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_HALF_NZERO"><span class="yiyi-st" id="yiyi-144"> <code class="descname">NPY_HALF_NZERO</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-145">此宏定义为负零。</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_HALF_ONE"><span class="yiyi-st" id="yiyi-146"> <code class="descname">NPY_HALF_ONE</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-147">此宏定义为1.0。</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_HALF_NEGONE"><span class="yiyi-st" id="yiyi-148"> <code class="descname">NPY_HALF_NEGONE</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-149">此宏定义为-1.0。</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_HALF_PINF"><span class="yiyi-st" id="yiyi-150"> <code class="descname">NPY_HALF_PINF</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-151">此宏定义为+ inf。</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_HALF_NINF"><span class="yiyi-st" id="yiyi-152"> <code class="descname">NPY_HALF_NINF</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-153">此宏定义为-inf。</span></p>
</dd></dl>
<dl class="var">
<dt id="c.NPY_HALF_NAN"><span class="yiyi-st" id="yiyi-154"> <code class="descname">NPY_HALF_NAN</code></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-155">此宏定义为NaN值，保证未设置其符号位。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_to_float"><span class="yiyi-st" id="yiyi-156"> float <code class="descname">npy_half_to_float</code><span class="sig-paren">(</span>npy_half<em> h</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-157">将半精度浮点数转换为单精度浮点数。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_to_double"><span class="yiyi-st" id="yiyi-158"> double <code class="descname">npy_half_to_double</code><span class="sig-paren">(</span>npy_half<em> h</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-159">将半精度浮点数转换为双精度浮点数。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_float_to_half"><span class="yiyi-st" id="yiyi-160"> npy_half <code class="descname">npy_float_to_half</code><span class="sig-paren">(</span>float<em> f</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-161">将单精度浮点数转换为半精度浮点数。</span><span class="yiyi-st" id="yiyi-162">该值四舍五入到最接近的可表示的一半，连接到最接近的偶数。</span><span class="yiyi-st" id="yiyi-163">如果值太小或太大，系统的浮点下溢或溢出位将被置1。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_double_to_half"><span class="yiyi-st" id="yiyi-164"> npy_half <code class="descname">npy_double_to_half</code><span class="sig-paren">(</span>double<em> d</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-165">将双精度浮点数转换为半精度浮点数。</span><span class="yiyi-st" id="yiyi-166">该值四舍五入到最接近的可表示的一半，连接到最接近的偶数。</span><span class="yiyi-st" id="yiyi-167">如果值太小或太大，系统的浮点下溢或溢出位将被置1。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_eq"><span class="yiyi-st" id="yiyi-168"> int <code class="descname">npy_half_eq</code><span class="sig-paren">(</span>npy_half<em> h1</em>, npy_half<em> h2</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-169">比较两个半精度浮点数（h1 == h2）。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_ne"><span class="yiyi-st" id="yiyi-170"> int <code class="descname">npy_half_ne</code><span class="sig-paren">(</span>npy_half<em> h1</em>, npy_half<em> h2</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-171">比较两个半精度浮点数（h1！= h2）。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_le"><span class="yiyi-st" id="yiyi-172"> int <code class="descname">npy_half_le</code><span class="sig-paren">(</span>npy_half<em> h1</em>, npy_half<em> h2</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-173">比较两个半精度浮点数（h1</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_lt"><span class="yiyi-st" id="yiyi-174"> int <code class="descname">npy_half_lt</code><span class="sig-paren">(</span>npy_half<em> h1</em>, npy_half<em> h2</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-175">比较两个半精度浮点数（h1</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_ge"><span class="yiyi-st" id="yiyi-176"> int <code class="descname">npy_half_ge</code><span class="sig-paren">(</span>npy_half<em> h1</em>, npy_half<em> h2</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-177">比较两个半精度浮点（h1> = h2）。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_gt"><span class="yiyi-st" id="yiyi-178"> int <code class="descname">npy_half_gt</code><span class="sig-paren">(</span>npy_half<em> h1</em>, npy_half<em> h2</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-179">比较两个半精度浮点数（h1> h2）。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_eq_nonan"><span class="yiyi-st" id="yiyi-180"> int <code class="descname">npy_half_eq_nonan</code><span class="sig-paren">(</span>npy_half<em> h1</em>, npy_half<em> h2</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-181">比较两个已知不是NaN（h1 == h2）的半精度浮点数。</span><span class="yiyi-st" id="yiyi-182">如果值为NaN，则结果未定义。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_lt_nonan"><span class="yiyi-st" id="yiyi-183"> int <code class="descname">npy_half_lt_nonan</code><span class="sig-paren">(</span>npy_half<em> h1</em>, npy_half<em> h2</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-184">比较两个已知不是NaN的半精度浮点数（h1</span><span class="yiyi-st" id="yiyi-185">如果值为NaN，则结果未定义。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_le_nonan"><span class="yiyi-st" id="yiyi-186"> int <code class="descname">npy_half_le_nonan</code><span class="sig-paren">(</span>npy_half<em> h1</em>, npy_half<em> h2</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-187">比较两个已知不是NaN的半精度浮点数（h1</span><span class="yiyi-st" id="yiyi-188">如果值为NaN，则结果未定义。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_iszero"><span class="yiyi-st" id="yiyi-189"> int <code class="descname">npy_half_iszero</code><span class="sig-paren">(</span>npy_half<em> h</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-190">测试半精度浮点是否具有等于零的值。</span><span class="yiyi-st" id="yiyi-191">这可能比调用npy_half_eq（h，NPY_ZERO）稍快。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_isnan"><span class="yiyi-st" id="yiyi-192"> int <code class="descname">npy_half_isnan</code><span class="sig-paren">(</span>npy_half<em> h</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-193">测试半精度浮点是否为NaN。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_isinf"><span class="yiyi-st" id="yiyi-194"> int <code class="descname">npy_half_isinf</code><span class="sig-paren">(</span>npy_half<em> h</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-195">测试半精度浮点是加或减Inf。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_isfinite"><span class="yiyi-st" id="yiyi-196"> int <code class="descname">npy_half_isfinite</code><span class="sig-paren">(</span>npy_half<em> h</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-197">测试半精度浮点是否是有限的（不是NaN或Inf）。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_signbit"><span class="yiyi-st" id="yiyi-198"> int <code class="descname">npy_half_signbit</code><span class="sig-paren">(</span>npy_half<em> h</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-199">返回1是h是负数，否则为0。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_copysign"><span class="yiyi-st" id="yiyi-200"> npy_half <code class="descname">npy_half_copysign</code><span class="sig-paren">(</span>npy_half<em> x</em>, npy_half<em> y</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-201">返回从y复制的符号位的x值。</span><span class="yiyi-st" id="yiyi-202">适用于任何值，包括Inf和NaN。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_spacing"><span class="yiyi-st" id="yiyi-203"> npy_half <code class="descname">npy_half_spacing</code><span class="sig-paren">(</span>npy_half<em> h</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-204">这与半精度浮点相同，如在低级浮点部分中描述的npy_spacing和npy_spacingf。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_half_nextafter"><span class="yiyi-st" id="yiyi-205"> npy_half <code class="descname">npy_half_nextafter</code><span class="sig-paren">(</span>npy_half<em> x</em>, npy_half<em> y</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-206">这与半精度浮点相同，与低级浮点部分中描述的npy_nextafter和npy_nextafterf相同。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_floatbits_to_halfbits"><span class="yiyi-st" id="yiyi-207"> npy_uint16 <code class="descname">npy_floatbits_to_halfbits</code><span class="sig-paren">(</span>npy_uint32<em> f</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-208">低级函数，将作为uint32存储的32位单精度浮点数转换为16位半精度浮点数。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_doublebits_to_halfbits"><span class="yiyi-st" id="yiyi-209"> npy_uint16 <code class="descname">npy_doublebits_to_halfbits</code><span class="sig-paren">(</span>npy_uint64<em> d</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-210">低级函数，将64位双精度浮点数转换为16位半精度浮点数，存储为uint64。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_halfbits_to_floatbits"><span class="yiyi-st" id="yiyi-211"> npy_uint32 <code class="descname">npy_halfbits_to_floatbits</code><span class="sig-paren">(</span>npy_uint16<em> h</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-212">低级函数，将16位半精度浮点数转换为32位单精度浮点数，存储为uint32。</span></p>
</dd></dl>
<dl class="function">
<dt id="c.npy_halfbits_to_doublebits"><span class="yiyi-st" id="yiyi-213"> npy_uint64 <code class="descname">npy_halfbits_to_doublebits</code><span class="sig-paren">(</span>npy_uint16<em> h</em><span class="sig-paren">)</span></span></dt>
<dd><p><span class="yiyi-st" id="yiyi-214">低级函数，将16位半精度浮点数转换为64位双精度浮点数，存储为uint64。</span></p>
</dd></dl>
</div>
</div>