1Stdlib.Bigarray(3) OCaml library Stdlib.Bigarray(3)
2
3
4
6 Stdlib.Bigarray - no description
7
9 Module Stdlib.Bigarray
10
12 Module Bigarray
13 : (module Stdlib__bigarray)
14
15
16
17
18
19
20
21
22
23 Element kinds
24 Bigarrays can contain elements of the following kinds:
25
26 -IEEE single precision (32 bits) floating-point numbers ( Bigar‐
27 ray.float32_elt ),
28
29 -IEEE double precision (64 bits) floating-point numbers ( Bigar‐
30 ray.float64_elt ),
31
32 -IEEE single precision (2 * 32 bits) floating-point complex numbers (
33 Bigarray.complex32_elt ),
34
35 -IEEE double precision (2 * 64 bits) floating-point complex numbers (
36 Bigarray.complex64_elt ),
37
38 -8-bit integers (signed or unsigned) ( Bigarray.int8_signed_elt or
39 Bigarray.int8_unsigned_elt ),
40
41 -16-bit integers (signed or unsigned) ( Bigarray.int16_signed_elt or
42 Bigarray.int16_unsigned_elt ),
43
44 -OCaml integers (signed, 31 bits on 32-bit architectures, 63 bits on
45 64-bit architectures) ( Bigarray.int_elt ),
46
47 -32-bit signed integers ( Bigarray.int32_elt ),
48
49 -64-bit signed integers ( Bigarray.int64_elt ),
50
51 -platform-native signed integers (32 bits on 32-bit architectures, 64
52 bits on 64-bit architectures) ( Bigarray.nativeint_elt ).
53
54 Each element kind is represented at the type level by one of the *_elt
55 types defined below (defined with a single constructor instead of
56 abstract types for technical injectivity reasons).
57
58 type float32_elt =
59 | Float32_elt
60
61
62
63
64 type float64_elt =
65 | Float64_elt
66
67
68
69
70 type int8_signed_elt =
71 | Int8_signed_elt
72
73
74
75
76 type int8_unsigned_elt =
77 | Int8_unsigned_elt
78
79
80
81
82 type int16_signed_elt =
83 | Int16_signed_elt
84
85
86
87
88 type int16_unsigned_elt =
89 | Int16_unsigned_elt
90
91
92
93
94 type int32_elt =
95 | Int32_elt
96
97
98
99
100 type int64_elt =
101 | Int64_elt
102
103
104
105
106 type int_elt =
107 | Int_elt
108
109
110
111
112 type nativeint_elt =
113 | Nativeint_elt
114
115
116
117
118 type complex32_elt =
119 | Complex32_elt
120
121
122
123
124 type complex64_elt =
125 | Complex64_elt
126
127
128
129
130 type ('a, 'b) kind =
131 | Float32 : (float, float32_elt) kind
132 | Float64 : (float, float64_elt) kind
133 | Int8_signed : (int, int8_signed_elt) kind
134 | Int8_unsigned : (int, int8_unsigned_elt) kind
135 | Int16_signed : (int, int16_signed_elt) kind
136 | Int16_unsigned : (int, int16_unsigned_elt) kind
137 | Int32 : (int32, int32_elt) kind
138 | Int64 : (int64, int64_elt) kind
139 | Int : (int, int_elt) kind
140 | Nativeint : (nativeint, nativeint_elt) kind
141 | Complex32 : (Complex.t, complex32_elt) kind
142 | Complex64 : (Complex.t, complex64_elt) kind
143 | Char : (char, int8_unsigned_elt) kind
144
145
146 To each element kind is associated an OCaml type, which is the type of
147 OCaml values that can be stored in the Bigarray or read back from it.
148 This type is not necessarily the same as the type of the array elements
149 proper: for instance, a Bigarray whose elements are of kind float32_elt
150 contains 32-bit single precision floats, but reading or writing one of
151 its elements from OCaml uses the OCaml type float , which is 64-bit
152 double precision floats.
153
154 The GADT type ('a, 'b) kind captures this association of an OCaml type
155 'a for values read or written in the Bigarray, and of an element kind
156 'b which represents the actual contents of the Bigarray. Its construc‐
157 tors list all possible associations of OCaml types with element kinds,
158 and are re-exported below for backward-compatibility reasons.
159
160 Using a generalized algebraic datatype (GADT) here allows writing
161 well-typed polymorphic functions whose return type depend on the argu‐
162 ment type, such as:
163
164
165 let zero : type a b. (a, b) kind -> a = function
166 | Float32 -> 0.0 | Complex32 -> Complex.zero
167 | Float64 -> 0.0 | Complex64 -> Complex.zero
168 | Int8_signed -> 0 | Int8_unsigned -> 0
169 | Int16_signed -> 0 | Int16_unsigned -> 0
170 | Int32 -> 0l | Int64 -> 0L
171 | Int -> 0 | Nativeint -> 0n
172 | Char -> '\000'
173
174
175
176
177 val float32 : (float, float32_elt) kind
178
179 See Bigarray.char .
180
181
182
183 val float64 : (float, float64_elt) kind
184
185 See Bigarray.char .
186
187
188
189 val complex32 : (Complex.t, complex32_elt) kind
190
191 See Bigarray.char .
192
193
194
195 val complex64 : (Complex.t, complex64_elt) kind
196
197 See Bigarray.char .
198
199
200
201 val int8_signed : (int, int8_signed_elt) kind
202
203 See Bigarray.char .
204
205
206
207 val int8_unsigned : (int, int8_unsigned_elt) kind
208
209 See Bigarray.char .
210
211
212
213 val int16_signed : (int, int16_signed_elt) kind
214
215 See Bigarray.char .
216
217
218
219 val int16_unsigned : (int, int16_unsigned_elt) kind
220
221 See Bigarray.char .
222
223
224
225 val int : (int, int_elt) kind
226
227 See Bigarray.char .
228
229
230
231 val int32 : (int32, int32_elt) kind
232
233 See Bigarray.char .
234
235
236
237 val int64 : (int64, int64_elt) kind
238
239 See Bigarray.char .
240
241
242
243 val nativeint : (nativeint, nativeint_elt) kind
244
245 See Bigarray.char .
246
247
248
249 val char : (char, int8_unsigned_elt) kind
250
251 As shown by the types of the values above, Bigarrays of kind
252 float32_elt and float64_elt are accessed using the OCaml type float .
253 Bigarrays of complex kinds complex32_elt , complex64_elt are accessed
254 with the OCaml type Complex.t . Bigarrays of integer kinds are accessed
255 using the smallest OCaml integer type large enough to represent the
256 array elements: int for 8- and 16-bit integer Bigarrays, as well as
257 OCaml-integer Bigarrays; int32 for 32-bit integer Bigarrays; int64 for
258 64-bit integer Bigarrays; and nativeint for platform-native integer
259 Bigarrays. Finally, Bigarrays of kind int8_unsigned_elt can also be
260 accessed as arrays of characters instead of arrays of small integers,
261 by using the kind value char instead of int8_unsigned .
262
263
264
265 val kind_size_in_bytes : ('a, 'b) kind -> int
266
267
268 kind_size_in_bytes k is the number of bytes used to store an element of
269 type k .
270
271
272 Since 4.03.0
273
274
275
276
277 Array layouts
278 type c_layout =
279 | C_layout_typ
280
281
282 See Bigarray.fortran_layout .
283
284
285 type fortran_layout =
286 | Fortran_layout_typ
287
288
289 To facilitate interoperability with existing C and Fortran code, this
290 library supports two different memory layouts for Bigarrays, one com‐
291 patible with the C conventions, the other compatible with the Fortran
292 conventions.
293
294 In the C-style layout, array indices start at 0, and multi-dimensional
295 arrays are laid out in row-major format. That is, for a two-dimen‐
296 sional array, all elements of row 0 are contiguous in memory, followed
297 by all elements of row 1, etc. In other terms, the array elements at
298 (x,y) and (x, y+1) are adjacent in memory.
299
300 In the Fortran-style layout, array indices start at 1, and multi-dimen‐
301 sional arrays are laid out in column-major format. That is, for a
302 two-dimensional array, all elements of column 0 are contiguous in mem‐
303 ory, followed by all elements of column 1, etc. In other terms, the
304 array elements at (x,y) and (x+1, y) are adjacent in memory.
305
306 Each layout style is identified at the type level by the phantom types
307 Bigarray.c_layout and Bigarray.fortran_layout respectively.
308
309
310
311
312 Supported layouts
313 The GADT type 'a layout represents one of the two supported memory lay‐
314 outs: C-style or Fortran-style. Its constructors are re-exported as
315 values below for backward-compatibility reasons.
316
317 type 'a layout =
318 | C_layout : c_layout layout
319 | Fortran_layout : fortran_layout layout
320
321
322
323
324
325 val c_layout : c_layout layout
326
327
328
329
330 val fortran_layout : fortran_layout layout
331
332
333
334
335
336 Generic arrays (of arbitrarily many dimensions)
337 module Genarray : sig end
338
339
340
341
342
343
344 Zero-dimensional arrays
345 module Array0 : sig end
346
347
348 Zero-dimensional arrays. The Array0 structure provides operations simi‐
349 lar to those of Bigarray.Genarray , but specialized to the case of
350 zero-dimensional arrays that only contain a single scalar value. Stat‐
351 ically knowing the number of dimensions of the array allows faster
352 operations, and more precise static type-checking.
353
354
355 Since 4.05.0
356
357
358
359
360 One-dimensional arrays
361 module Array1 : sig end
362
363
364 One-dimensional arrays. The Array1 structure provides operations simi‐
365 lar to those of Bigarray.Genarray , but specialized to the case of
366 one-dimensional arrays. (The Bigarray.Array2 and Bigarray.Array3
367 structures below provide operations specialized for two- and
368 three-dimensional arrays.) Statically knowing the number of dimensions
369 of the array allows faster operations, and more precise static
370 type-checking.
371
372
373
374
375 Two-dimensional arrays
376 module Array2 : sig end
377
378
379 Two-dimensional arrays. The Array2 structure provides operations simi‐
380 lar to those of Bigarray.Genarray , but specialized to the case of
381 two-dimensional arrays.
382
383
384
385
386 Three-dimensional arrays
387 module Array3 : sig end
388
389
390 Three-dimensional arrays. The Array3 structure provides operations sim‐
391 ilar to those of Bigarray.Genarray , but specialized to the case of
392 three-dimensional arrays.
393
394
395
396
397 Coercions between generic Bigarrays and fixed-dimension Bigarrays
398 val genarray_of_array0 : ('a, 'b, 'c) Array0.t -> ('a, 'b, 'c) Genar‐
399 ray.t
400
401 Return the generic Bigarray corresponding to the given zero-dimensional
402 Bigarray.
403
404
405 Since 4.05.0
406
407
408
409 val genarray_of_array1 : ('a, 'b, 'c) Array1.t -> ('a, 'b, 'c) Genar‐
410 ray.t
411
412 Return the generic Bigarray corresponding to the given one-dimensional
413 Bigarray.
414
415
416
417 val genarray_of_array2 : ('a, 'b, 'c) Array2.t -> ('a, 'b, 'c) Genar‐
418 ray.t
419
420 Return the generic Bigarray corresponding to the given two-dimensional
421 Bigarray.
422
423
424
425 val genarray_of_array3 : ('a, 'b, 'c) Array3.t -> ('a, 'b, 'c) Genar‐
426 ray.t
427
428 Return the generic Bigarray corresponding to the given three-dimen‐
429 sional Bigarray.
430
431
432
433 val array0_of_genarray : ('a, 'b, 'c) Genarray.t -> ('a, 'b, 'c)
434 Array0.t
435
436 Return the zero-dimensional Bigarray corresponding to the given generic
437 Bigarray.
438
439
440 Since 4.05.0
441
442
443 Raises Invalid_argument if the generic Bigarray does not have exactly
444 zero dimension.
445
446
447
448 val array1_of_genarray : ('a, 'b, 'c) Genarray.t -> ('a, 'b, 'c)
449 Array1.t
450
451 Return the one-dimensional Bigarray corresponding to the given generic
452 Bigarray.
453
454
455 Raises Invalid_argument if the generic Bigarray does not have exactly
456 one dimension.
457
458
459
460 val array2_of_genarray : ('a, 'b, 'c) Genarray.t -> ('a, 'b, 'c)
461 Array2.t
462
463 Return the two-dimensional Bigarray corresponding to the given generic
464 Bigarray.
465
466
467 Raises Invalid_argument if the generic Bigarray does not have exactly
468 two dimensions.
469
470
471
472 val array3_of_genarray : ('a, 'b, 'c) Genarray.t -> ('a, 'b, 'c)
473 Array3.t
474
475 Return the three-dimensional Bigarray corresponding to the given
476 generic Bigarray.
477
478
479 Raises Invalid_argument if the generic Bigarray does not have exactly
480 three dimensions.
481
482
483
484
485 Re-shaping Bigarrays
486 val reshape : ('a, 'b, 'c) Genarray.t -> int array -> ('a, 'b, 'c)
487 Genarray.t
488
489
490 reshape b [|d1;...;dN|] converts the Bigarray b to a N -dimensional
491 array of dimensions d1 ... dN . The returned array and the original
492 array b share their data and have the same layout. For instance,
493 assuming that b is a one-dimensional array of dimension 12, reshape b
494 [|3;4|] returns a two-dimensional array b' of dimensions 3 and 4. If b
495 has C layout, the element (x,y) of b' corresponds to the element x * 3
496 + y of b . If b has Fortran layout, the element (x,y) of b' corre‐
497 sponds to the element x + (y - 1) * 4 of b . The returned Bigarray
498 must have exactly the same number of elements as the original Bigarray
499 b . That is, the product of the dimensions of b must be equal to i1 *
500 ... * iN . Otherwise, Invalid_argument is raised.
501
502
503
504 val reshape_0 : ('a, 'b, 'c) Genarray.t -> ('a, 'b, 'c) Array0.t
505
506 Specialized version of Bigarray.reshape for reshaping to zero-dimen‐
507 sional arrays.
508
509
510 Since 4.05.0
511
512
513
514 val reshape_1 : ('a, 'b, 'c) Genarray.t -> int -> ('a, 'b, 'c) Array1.t
515
516 Specialized version of Bigarray.reshape for reshaping to one-dimen‐
517 sional arrays.
518
519
520
521 val reshape_2 : ('a, 'b, 'c) Genarray.t -> int -> int -> ('a, 'b, 'c)
522 Array2.t
523
524 Specialized version of Bigarray.reshape for reshaping to two-dimen‐
525 sional arrays.
526
527
528
529 val reshape_3 : ('a, 'b, 'c) Genarray.t -> int -> int -> int -> ('a,
530 'b, 'c) Array3.t
531
532 Specialized version of Bigarray.reshape for reshaping to three-dimen‐
533 sional arrays.
534
535
536
537
538
539OCamldoc 2020-09-01 Stdlib.Bigarray(3)