Template library for multi-GPU pseudorandom number recursion-based generators

Author

Szalkowski, Dominik ; Stpiczynski, Przemyslaw

Author_Institution

Inst. of Math., Maria Curie-Sklodowska Univ., Lublin, Poland

fYear

2013

fDate

8-11 Sept. 2013

Firstpage

515

Lastpage

519

Abstract

The aim of the paper is to show how to design and implement fast parallel algorithms for Linear Congruential, Lagged Fibonacci and Wichmann-Hill pseudorandom number generators. The new algorithms employ the divide-and-conquer approach for solving linear recurrence systems. They are implemented on multi GPU-accelerated systems using CUDA. Numerical experiments performed on a computer system with two Fermi GPU cards show that our software achieve good performance in comparison to the widely used NVIDIA CURAND Library.

Keywords

graphics processing units; parallel algorithms; parallel architectures; random number generation; CUDA; Fermi GPU cards; NVIDIA CURAND library; Wichmann-Hill pseudorandom number generators; compute unified device architecture; fast parallel algorithms; graphics processing unit; lagged Fibonacci pseudorandom number generators; linear congruential pseudorandom number generators; linear recurrence systems; multiGPU pseudorandom number recursion-based generators; multiGPU-accelerated systems; Computers; Generators; Graphics processing units; Libraries; Parallel algorithms; Performance evaluation; Vectors;

fLanguage

English

Publisher

ieee

Conference_Titel

Computer Science and Information Systems (FedCSIS), 2013 Federated Conference on

Conference_Location

Krako??w

Type

conf

Filename

6644049