Computer Software
graphics processing units
generator
CPU
statistical quality
Tyche
parallel streams
false
independent streams
https://scigraph.springernature.com/explorer/license/
function
2012
computer simulations
2022-12-01T06:51
simulations
iteration function
Fast and Small Nonlinear Pseudorandom Number Generators for Computer Simulation
chapter
pseudorandom numbers
number
parallel environment
performance
state
nonlinear generators
parallelism
processing unit
overlap
number generator
nonlinear pseudorandom number generators
excellent statistical properties
today
ChaCha
architecture
pseudorandom number generators
https://doi.org/10.1007/978-3-642-31464-3_10
statistical properties
2012-01-01
cipher
properties
quality
parallel independent streams
length
units
bounds
stream cipher
example
paper
environment
high performance
comparison
fast stream ciphers
large number
streams
chapters
period length
In this paper we present Tyche, a nonlinear pseudorandom number generator designed for computer simulation. Tyche has a small 128-bit state and an expected period length of 2127. Unlike most nonlinear generators, Tyche is consistently fast across architectures, due to its very simple iteration function derived from ChaCha, one of today’s fastest stream ciphers.Tyche is especially amenable for the highly parallel environments we find today, in particular for Graphics Processing Units (GPUs), where it enables a very large number of uncorrelated parallel streams running independently. For example, 216 parallel independent streams are expected to generate about 296 pseudorandom numbers each, without overlaps.Additionally, we determine bounds for the period length and parallelism of our generators, and evaluate their statistical quality and performance. We compare Tyche and the variant Tyche-i to the XORWOW and TEA8 generators in CPUs and GPUs. Our comparisons show that Tyche and Tyche-i simultaneously achieve high performance and excellent statistical properties, particularly when compared to other nonlinear generators.
92-101
Information and Computing Sciences
CISUC, Dept. of Informatics Engineering, University of Coimbra, Portugal
CISUC, Dept. of Informatics Engineering, University of Coimbra, Portugal
doi
10.1007/978-3-642-31464-3_10
Filipe
Araujo
Springer Nature - SN SciGraph project
Neves
Samuel
Springer Nature
Konrad
Karczewski
Dongarra
Jack
Waśniewski
Jerzy
978-3-642-31463-6
978-3-642-31464-3
Parallel Processing and Applied Mathematics
Roman
Wyrzykowski
dimensions_id
pub.1029031748