GPU accelerated population annealing algorithm

DOI10.1016/J.CPC.2017.06.020MaRDI QIDQ1737463zbMATH OpenOpenAlexFDO

Authors Lev Yu. Barash, Martin Weigel, Michal Borovský, Wolfhard Janke, L. Shchur

Publication date 23 April 2019

Published in Computer Physics Communications (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1703.03676

zbMATH Keywords

graphics processing unitsIsing modelMonte Carlo simulationsmulti-spin codingparallel computingpopulation annealing

Mathematics Subject Classification ID

Monte Carlo methods (65C05) Software, source code, etc. for problems pertaining to statistical mechanics (82-04)

Abstract: Population annealing is a promising recent approach for Monte Carlo simulations in statistical physics, in particular for the simulation of systems with complex free-energy landscapes. It is a hybrid method, combining importance sampling through Markov chains with elements of sequential Monte Carlo in the form of population control. While it appears to provide algorithmic capabilities for the simulation of such systems that are roughly comparable to those of more established approaches such as parallel tempering, it is intrinsically much more suitable for massively parallel computing. Here, we tap into this structural advantage and present a highly optimized implementation of the population annealing algorithm on GPUs that promises speed-ups of several orders of magnitude as compared to a serial implementation on CPUs. While the sample code is for simulations of the 2D ferromagnetic Ising model, it should be easily adapted for simulations of other spin models, including disordered systems. Our code includes implementations of some advanced algorithmic features that have only recently been suggested, namely the automatic adaptation of temperature steps and a multi-histogram analysis of the data at different temperatures.

Recommendations

• Multi-GPU accelerated multi-spin Monte Carlo simulations of the 2D Ising model
• GPU accelerated Monte Carlo simulation of the 2D and 3D Ising model
• An efficient implementation of parallel simulated annealing algorithm in GPUs
• Performance potential for simulating spin models on GPU
• scientific article; zbMATH DE number 6985549

Cites work

• A guide to Monte Carlo simulations in statistical physics.
• Go with the winners: a general Monte Carlo strategy
• GPU accelerated Monte Carlo simulation of the 2D and 3D Ising model
• scientific article; zbMATH DE number 5050923 (Why is no real title available?)
• Massively parallelized replica-exchange simulations of polymers on GPUs
• Numerical recipes. The art of scientific computing.
• Performance potential for simulating spin models on GPU
• Sequential Monte Carlo Methods in Practice
• TestU01
• The critical behavior of 3D Ising spin glass models: universality and scaling corrections

Cited in

(7)

• An efficient implementation of parallel simulated annealing algorithm in GPUs
• Parallelism of the finite-time dynamics method based on GPU
• Model-free data-driven inference in computational mechanics
• PAIsing
• Massively parallel simulations for disordered systems
• Population Monte Carlo as a meta-algorithm
• High-performance statistical computing (HPSC): challenges, opportunities, and future directions

Describes a project that uses

Uses Software

• TestU01
• CUDA
• Janus
• PRAND

This page was built for publication: GPU accelerated population annealing algorithm

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1737463)