Skip to content
/ gpusim Public

Code for ITNG2019 paper ( Best student paper award ): A graphical processing unit accelerated NORmal To Anything algorithm for high dimensional multivariate simulation. Published in Advances in Intellegent Systems and Computing Volume 800

Notifications You must be signed in to change notification settings

xliUNR/gpusim

Repository files navigation

gpusim

A graphical processing unit accelerated NOrmal To Anything algorithm for high dimensional multivariate simulation

Project for CS791: GPU programming.
This project implements NORTA transformations for generating random variables with given correlation matrix.
Presented at ITNG 2019. Won best student paper award.

Link to publication: https://link.springer.com/chapter/10.1007/978-3-030-14070-0_46

How to use program:

makefile is in build folder.
Executable is called gpusim
Command line arguments are:
<correlation matrix file path> <distributions file path> d n seed <output file name> runtype inverseflag

d is the number of random variables in random vector (also dimension of corr matrix )
n is the number of simulation replicates
seed is the seed used for random number generator
output file name is the file name of the simulation matrix file that will be generated at the end of program

runtype is an integer. 0 for sequential run, 1 for parallel GPU.
inverseflag is an integer. 0 to not run inverse transform, 1 to run inverse transform.


There is a space between each argument.

call goes like this:
gpusim ../test_corr_matrix_d=20501.txt ../betar20501.csv 20501 2000 154224 simOutputFile 0 0

About

Code for ITNG2019 paper ( Best student paper award ): A graphical processing unit accelerated NORmal To Anything algorithm for high dimensional multivariate simulation. Published in Advances in Intellegent Systems and Computing Volume 800

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published