Reference Parallel Primitives#

These are reference implementations of basic parallel primitives. These are not actually parallel, but work the same way. They are good for testing.

sage.parallel.reference.parallel_iter(f, inputs)[source]#

Reference parallel iterator implementation.

INPUT:

f – a Python function that can be pickled using the pickle_function command.
inputs – a list of pickleable pairs (args, kwds), where args is a tuple and kwds is a dictionary.

OUTPUT:

iterator over 2-tuples (inputs[i], f(inputs[i])), where the order may be completely random

EXAMPLES:

Sage

sage: def f(N,M=10): return N*M
sage: inputs = [((2,3),{}),  (tuple(), {'M':5,'N':3}), ((2,),{})]
sage: set_random_seed(0)
sage: for a, val in sage.parallel.reference.parallel_iter(f, inputs):
....:     print((a, val))
(((2,), {}), 20)
(((), {'M': 5, 'N': 3}), 15)
(((2, 3), {}), 6)
sage: for a, val in sage.parallel.reference.parallel_iter(f, inputs):
....:     print((a, val))
(((), {'M': 5, 'N': 3}), 15)
(((2,), {}), 20)
(((2, 3), {}), 6)

Python

>>> from sage.all import *
>>> def f(N,M=Integer(10)): return N*M
>>> inputs = [((Integer(2),Integer(3)),{}),  (tuple(), {'M':Integer(5),'N':Integer(3)}), ((Integer(2),),{})]
>>> set_random_seed(Integer(0))
>>> for a, val in sage.parallel.reference.parallel_iter(f, inputs):
...     print((a, val))
(((2,), {}), 20)
(((), {'M': 5, 'N': 3}), 15)
(((2, 3), {}), 6)
>>> for a, val in sage.parallel.reference.parallel_iter(f, inputs):
...     print((a, val))
(((), {'M': 5, 'N': 3}), 15)
(((2,), {}), 20)
(((2, 3), {}), 6)