Treffer: clrng: a tool set for parallel random number generation on GPUs in R.

{A novel GPU-accelerated package is proposed to enable efficient parallel random number generation in R, significantly improving performance for large-scale statistical simulations.}<bold>(i) Context</bold>: Parallel processing with Graphics Processing Units (GPUs) can speed up computationally intensive tasks, which when combined with R, it can largely improve R's limitations in terms of speed, memory usage and single-threaded computation. <bold>(ii) Problem</bold>: Despite the importance of random number generation for simulation-based statistical inference and modelling, there is currently no R package that supports reproducible, GPU-based parallel random number generation. <bold>(iii) Solution</bold>: To fill this gap, we introduce the R package <bold>clrng</bold>, which integrates the OpenCL <bold>clRNG</bold> library with the <bold>gpuR</bold> package to enable efficient parallel random number generation on GPUs. <bold>(iv) Results</bold>: <bold>clrng</bold> enables reproducible research by setting random initial seeds for streams on both GPU and CPU, thereby accelerating the performance of several types of statistical simulation and modelling. The random number generator in <bold>clrng</bold> guarantees independent parallel samples even in interactive, ad-hoc R sessions (e.g. interrupted and resumed). This package is portable and flexible, allowing developers to embed its random number generation kernel into a wide range of statistical applications. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Statistical Computation & Simulation is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)