OpenDX - Documentation
Full Contents QuickStart Guide User's Guide User's Reference
Previous Page Next Page Table of Contents Partial Table of Contents Index Search

Partition

Category

Import and Export

Function

Partitions a data set for parallel processing.

Syntax


partitioned = Partition(input, n, size);

Inputs
Name Type Default Description
input field or group none field to be partitioned
n integer machine dependent maximum number of subparts
size integer one primitive threshold for partitioning

Outputs
Name Type Description
partitioned field or group set of partitioned fields

Functional Details

This module partitions a data set for parallel processing on an SMP multiprocessor machine. (You must be using Data Explorer SMP to take advantage of this feature.) Its output is a composite field, which is treated as a single entity by other modules.

input

is the input object to be partitioned. If this parameter specifies a group, each group member is partitioned with the same n and size parameters.

n

is the approximate number of partitions to be created. However, the module will not create partitions smaller than size (see below).

size

is the minimum number of connection elements per partition.

If n × size is larger than the total number of points, the output number of partitions may be smaller than n

If you do not specify n or size, appropriate default values are supplied, depending on the number of processors available.

Note: On a uniprocessor machine, n = 1 by default. Consequently, the same programs can be run on uniprocessor and multiprocessor machines without modification.

Components

All components in the input are propagated to the output.

Example Visual Programs

ExampleSMP.net

See Also

 Import


Full Contents QuickStart Guide User's Guide User's Reference

[ OpenDX Home at IBM | OpenDX.org ]