SystemML Class
std/2009/random/numeric
Summary
Source of pseudo-random numeric data
Status
Stable

Overview

A source of pseudo-random numeric data. The data output is generated at run-time by an internal random number generator, and is repeatable by seeding with an explicit Seed.

Notes

This process uses std/util/rng internally - processes only needing random data internally should use this utility, rather than accepting input from this process.

Connectivity

  • No inputs.
  • One output, the data generated.

State

This Process stores its State in DataML.

dims (INT64 1xN)
The dimension of the output data as [a, b, ...].
dist (STRING)
Distribution, one of "normal", "uniform" or "exponential".
pars (DOUBLE 1xP)
Specific to distribution. May be [mean sd] (normal), [min max] (uniform, value of exactly "max" is never output) or [min sd] (exponential).
complex (BOOLEAN scalar) OPTIONAL
If present and true, generate complex output data (i.e. twice as much output).
outputName (STRING) OPTIONAL
If supplied, the name of the single output (if not, "out" is used).

Example

Script

M Source Code (against 995)
fS = 1000; sys = sml_system; state = []; state.dims = [3 2]; state.dist = 'normal'; state.pars = [3 0.1]; sys = sys.addprocess('src', 'std/2009/random/numeric', fS, state); exe = brahms_execution; exe.all = true; exe.stop = 1; out = brahms(sys, exe); out.src plot(reshape(out.src.out, 6000, []))

Expected Output

Matlab Console
ans = out: [3x2x1000 double]