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
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
|
import sys
import itertools
from mpi4py import MPI
try:
import array
except ImportError:
array = None
try:
import numpy
except ImportError:
numpy = None
try:
import cupy
cupy_version = tuple(map(int, cupy.__version__.split('.', 2)[:2]))
except ImportError:
cupy = None
try:
import numba
import numba.cuda
numba_version = tuple(map(int, numba.__version__.split('.', 2)[:2]))
if numba_version < (0, 48):
import warnings
try:
warnings.warn(
'To test Numba GPU arrays, use Numba v0.48.0+.',
RuntimeWarning, stacklevel=1,
)
except RuntimeWarning:
pass
del numba_version
numba = None
except ImportError:
numba = None
__all__ = ['loop', 'test']
def make_typemap(entries):
if sys.version_info[:2] > (3, 7):
dict_type = dict
else:
from collections import OrderedDict
dict_type = OrderedDict
typemap = dict_type(
(typecode, datatype)
for typecode, datatype in entries
if datatype != MPI.DATATYPE_NULL
)
return typemap
TypeMap = make_typemap([
('b', MPI.SIGNED_CHAR),
('h', MPI.SHORT),
('i', MPI.INT),
('l', MPI.LONG),
('q', MPI.LONG_LONG),
('f', MPI.FLOAT),
('d', MPI.DOUBLE),
('g', MPI.LONG_DOUBLE),
])
TypeMapBool = make_typemap([
('?', MPI.C_BOOL),
])
TypeMapInteger = make_typemap([
('b', MPI.SIGNED_CHAR),
('h', MPI.SHORT),
('i', MPI.INT),
('l', MPI.LONG),
('q', MPI.LONG_LONG),
])
TypeMapUnsigned = make_typemap([
('B', MPI.UNSIGNED_CHAR),
('H', MPI.UNSIGNED_SHORT),
('I', MPI.UNSIGNED_INT),
('L', MPI.UNSIGNED_LONG),
('Q', MPI.UNSIGNED_LONG_LONG),
])
TypeMapFloat = make_typemap([
('f', MPI.FLOAT),
('d', MPI.DOUBLE),
('g', MPI.LONG_DOUBLE),
])
TypeMapComplex = make_typemap([
('F', MPI.C_FLOAT_COMPLEX),
('D', MPI.C_DOUBLE_COMPLEX),
('G', MPI.C_LONG_DOUBLE_COMPLEX),
])
ArrayBackends = []
def add_backend(cls):
ArrayBackends.append(cls)
return cls
class BaseArray:
backend = None
TypeMap = TypeMap.copy()
TypeMap.pop('g', None)
def __len__(self):
return len(self.array)
def __getitem__(self, i):
return self.array[i]
def __setitem__(self, i, v):
self.array[i] = v
@property
def mpidtype(self):
try:
return self.TypeMap[self.typecode]
except KeyError:
return MPI.DATATYPE_NULL
def as_raw(self):
return self.array
def as_mpi(self):
return (self.as_raw(), self.mpidtype)
def as_mpi_c(self, count):
return (self.as_raw(), count, self.mpidtype)
def as_mpi_v(self, cnt, dsp):
return (self.as_raw(), (cnt, dsp), self.mpidtype)
if array is not None:
def product(seq):
res = 1
for s in seq:
res = res * s
return res
def mkshape(shape):
return tuple([int(s) for s in shape])
@add_backend
class ArrayArray(BaseArray):
backend = 'array'
def __init__(self, arg, typecode, shape=None):
if isinstance(arg, (int, float)):
if shape is None:
shape = ()
else:
try:
shape = mkshape(shape)
except TypeError:
shape = (int(shape),)
size = product(shape)
arg = [arg] * size
else:
size = len(arg)
if shape is None:
shape = (size,)
else:
shape = mkshape(shape)
assert size == product(shape)
self.array = array.array(typecode, arg)
@property
def address(self):
return self.array.buffer_info()[0]
@property
def typecode(self):
return self.array.typecode
@property
def itemsize(self):
return self.array.itemsize
@property
def flat(self):
return self.array
@property
def size(self):
return self.array.buffer_info()[1]
if numpy is not None:
@add_backend
class ArrayNumPy(BaseArray):
backend = 'numpy'
TypeMap = make_typemap([])
TypeMap.update(TypeMapBool)
TypeMap.update(TypeMapInteger)
TypeMap.update(TypeMapUnsigned)
TypeMap.update(TypeMapFloat)
TypeMap.update(TypeMapComplex)
def __init__(self, arg, typecode, shape=None):
if isinstance(arg, (int, float, complex)):
if shape is None:
shape = ()
else:
if shape is None:
shape = len(arg)
self.array = numpy.zeros(shape, typecode)
if isinstance(arg, (int, float, complex)):
arg = numpy.asarray(arg).astype(typecode)
self.array.fill(arg)
else:
arg = numpy.asarray(arg).astype(typecode)
self.array[...] = arg
@property
def address(self):
return self.array.__array_interface__['data'][0]
@property
def typecode(self):
return self.array.dtype.char
@property
def itemsize(self):
return self.array.itemsize
@property
def flat(self):
return self.array.flat
@property
def size(self):
return self.array.size
try:
import dlpackimpl as dlpack
except ImportError:
dlpack = None
class BaseDLPackCPU:
def __dlpack_device__(self):
return (dlpack.DLDeviceType.kDLCPU, 0)
def __dlpack__(self, stream=None):
assert stream is None
capsule = dlpack.make_py_capsule(self.array)
return capsule
def as_raw(self):
return self
if dlpack is not None and array is not None:
@add_backend
class DLPackArray(BaseDLPackCPU, ArrayArray):
backend = 'dlpack-array'
def __init__(self, arg, typecode, shape=None):
super().__init__(arg, typecode, shape)
if dlpack is not None and numpy is not None:
@add_backend
class DLPackNumPy(BaseDLPackCPU, ArrayNumPy):
backend = 'dlpack-numpy'
def __init__(self, arg, typecode, shape=None):
super().__init__(arg, typecode, shape)
def typestr(typecode, itemsize):
typestr = ''
if sys.byteorder == 'little':
typestr += '<'
if sys.byteorder == 'big':
typestr += '>'
if typecode in '?':
typestr += 'b'
if typecode in 'bhilq':
typestr += 'i'
if typecode in 'BHILQ':
typestr += 'u'
if typecode in 'fdg':
typestr += 'f'
if typecode in 'FDG':
typestr += 'c'
typestr += str(itemsize)
return typestr
class BaseFakeGPUArray:
def set_interface(self, shape, readonly=False):
self.__cuda_array_interface__ = dict(
version = 0,
data = (self.address, readonly),
typestr = typestr(self.typecode, self.itemsize),
shape = shape,
)
def as_raw(self):
return self
if array is not None:
@add_backend
class FakeGPUArrayBasic(BaseFakeGPUArray, ArrayArray):
def __init__(self, arg, typecode, shape=None, readonly=False):
super().__init__(arg, typecode, shape)
self.set_interface((len(self),), readonly)
if numpy is not None:
@add_backend
class FakeGPUArrayNumPy(BaseFakeGPUArray, ArrayNumPy):
def __init__(self, arg, typecode, shape=None, readonly=False):
super().__init__(arg, typecode, shape)
self.set_interface(self.array.shape, readonly)
if cupy is not None:
@add_backend
class GPUArrayCuPy(BaseArray):
backend = 'cupy'
TypeMap = make_typemap([])
if cupy_version >= (11, 6):
TypeMap.update(TypeMapBool)
TypeMap.update(TypeMapInteger)
TypeMap.update(TypeMapUnsigned)
TypeMap.update(TypeMapFloat)
TypeMap.update(TypeMapComplex)
try:
cupy.array(0, 'g')
except ValueError:
TypeMap.pop('g', None)
try:
cupy.array(0, 'G')
except ValueError:
TypeMap.pop('G', None)
def __init__(self, arg, typecode, shape=None, readonly=False):
if isinstance(arg, (int, float, complex)):
if shape is None:
shape = ()
else:
if shape is None:
shape = len(arg)
self.array = cupy.zeros(shape, typecode)
if isinstance(arg, (int, float, complex)):
self.array.fill(arg)
else:
self.array[:] = cupy.asarray(arg, typecode)
@property
def address(self):
return self.array.__cuda_array_interface__['data'][0]
@property
def typecode(self):
return self.array.dtype.char
@property
def itemsize(self):
return self.array.itemsize
@property
def flat(self):
return self.array.ravel()
@property
def size(self):
return self.array.size
def as_raw(self):
cupy.cuda.get_current_stream().synchronize()
return self.array
if cupy is not None:
# Note: we do not create a BaseDLPackGPU class because each GPU library
# has its own way to get device ID etc, so we have to reimplement the
# DLPack support anyway
@add_backend
class DLPackCuPy(GPUArrayCuPy):
backend = 'dlpack-cupy'
has_dlpack = None
dev_type = None
def __init__(self, arg, typecode, shape=None):
super().__init__(arg, typecode, shape)
self.has_dlpack = hasattr(self.array, '__dlpack_device__')
# TODO(leofang): test CUDA managed memory?
if cupy.cuda.runtime.is_hip:
self.dev_type = dlpack.DLDeviceType.kDLROCM
else:
self.dev_type = dlpack.DLDeviceType.kDLCUDA
def __dlpack_device__(self):
if self.has_dlpack:
return self.array.__dlpack_device__()
else:
return (self.dev_type, self.array.device.id)
def __dlpack__(self, stream=None):
cupy.cuda.get_current_stream().synchronize()
if self.has_dlpack:
return self.array.__dlpack__(stream=-1)
else:
return self.array.toDlpack()
def as_raw(self):
return self
if numba is not None:
@add_backend
class GPUArrayNumba(BaseArray):
backend = 'numba'
TypeMap = make_typemap([])
TypeMap.update(TypeMapBool)
TypeMap.update(TypeMapInteger)
TypeMap.update(TypeMapUnsigned)
TypeMap.update(TypeMapFloat)
TypeMap.update(TypeMapComplex)
# one can allocate arrays with those types,
# but the Numba compiler doesn't support them...
TypeMap.pop('g', None)
TypeMap.pop('G', None)
def __init__(self, arg, typecode, shape=None, readonly=False):
if isinstance(arg, (int, float, complex)):
if shape is None:
shape = ()
else:
if shape is None:
shape = len(arg)
self.array = numba.cuda.device_array(shape, typecode)
if isinstance(arg, (int, float, complex)):
if self.array.size > 0:
self.array[:] = arg
elif arg == [] or arg == ():
self.array = numba.cuda.device_array(0, typecode)
else:
if self.array.size > 0:
self.array[:] = numba.cuda.to_device(arg)
# def __getitem__(self, i):
# if isinstance(i, slice):
# return self.array[i]
# elif i < self.array.size:
# return self.array[i]
# else:
# raise StopIteration
@property
def address(self):
return self.array.__cuda_array_interface__['data'][0]
@property
def typecode(self):
return self.array.dtype.char
@property
def itemsize(self):
return self.array.dtype.itemsize
@property
def flat(self):
if self.array.ndim <= 1:
return self.array
else:
return self.array.ravel()
@property
def size(self):
return self.array.size
def as_raw(self):
# numba by default always runs on the legacy default stream
numba.cuda.default_stream().synchronize()
return self.array
def loop(*args):
loop.array = None
loop.typecode = None
for array in ArrayBackends:
loop.array = array
for typecode in array.TypeMap:
loop.typecode = typecode
if not args:
yield array, typecode
else:
for prod in itertools.product(*args):
yield (array, typecode) + prod
del loop.array
del loop.typecode
def test(case, **kargs):
return case.subTest(
typecode=loop.typecode,
backend=loop.array.backend,
**kargs,
)
def scalar(arg):
return loop.array(arg, loop.typecode, 1)[0]
|
