We present VegasFlow, a new software for fast evaluation of high dimensional integrals based on Monte Carlo integration techniques designed for platforms with hardware accelerators. The growing complexity of calculations and simulations in many areas of science have been accompanied by advances in the computational tools which have helped their developments. VegasFlow enables developers to delegate all complicated aspects of hardware or platform implementation to the library so they can focus on the problem at hand. This software is inspired on the Vegas algorithm, ubiquitous in the particle physics community as the driver of cross section integration, and based on Google's powerful TensorFlow library. We benchmark the performance of this library on many different consumer and professional grade GPUs and CPUs. Program summary: Program Title: VegasFlow CPC Library link to program files: http://dx.doi.org.pros.lib.unimi.it/10.17632/rpgcbzzhdt.1 Developer's repository link: https://github.com/N3PDF/vegasflow Licensing provisions: GPLv3 Programming language: Python Nature of problem: The solution of high dimensional integrals requires the implementation of Monte Carlo algorithms such as Vegas. Monte Carlo algorithms are known to require long computation times. Solution method: Implementation of the Vegas algorithm using the dataflow graph infrastructure provided by the TensorFlow framework. Extension of the algorithm to take advantage of multi-threading CPU and multi-GPU setups.
VegasFlow: Accelerating Monte Carlo simulation across multiple hardware platforms / S. Carrazza, C. Juan. - In: COMPUTER PHYSICS COMMUNICATIONS. - ISSN 0010-4655. - 254(2020 Sep).
|Titolo:||VegasFlow: Accelerating Monte Carlo simulation across multiple hardware platforms|
CARRAZZA, STEFANO (Primo) (Corresponding)
CRUZ MARTINEZ, JUAN MANUEL (Ultimo)
|Parole Chiave:||Graphs; Hardware acceleration; Integration; Machine learning; Monte Carlo;|
|Settore Scientifico Disciplinare:||Settore FIS/02 - Fisica Teorica, Modelli e Metodi Matematici|
|Data di pubblicazione:||set-2020|
|Data ahead of print / Data di stampa:||16-mag-2020|
|Digital Object Identifier (DOI):||http://dx.doi.org/10.1016/j.cpc.2020.107376|
|Appare nelle tipologie:||01 - Articolo su periodico|