Performance potential for simulating spin models on GPU

DOI10.1016/J.JCP.2011.12.008zbMATH Open1245.82035arXiv1101.1427OpenAlexW2168891527MaRDI QIDQ417832FDOQ417832

Martin Weigel

Publication date: 14 May 2012

Published in: Journal of Computational Physics (Search for Journal in Brave)

Abstract: Graphics processing units (GPUs) are recently being used to an increasing degree for general computational purposes. This development is motivated by their theoretical peak performance, which significantly exceeds that of broadly available CPUs. For practical purposes, however, it is far from clear how much of this theoretical performance can be realized in actual scientific applications. As is discussed here for the case of studying classical spin models of statistical mechanics by Monte Carlo simulations, only an explicit tailoring of the involved algorithms to the specific architecture under consideration allows to harvest the computational power of GPU systems. A number of examples, ranging from Metropolis simulations of ferromagnetic Ising models, over continuous Heisenberg and disordered spin-glass systems to parallel-tempering simulations are discussed. Significant speed-ups by factors of up to 1000 compared to serial CPU code as well as previous GPU implementations are observed.

Full work available at URL: https://arxiv.org/abs/1101.1427

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
• Benchmarking GPU and CPU codes for Heisenberg spin glass over-relaxation
• GPU-based single-cluster algorithm for the simulation of the Ising model
• scientific article; zbMATH DE number 6985549

zbMATH Keywords

parallel temperingIsing modelMonte Carlo simulationsspin glassesgraphics processing unitsHeisenberg model

Mathematics Subject Classification ID

Lattice systems (Ising, dimer, Potts, etc.) and systems on graphs arising in equilibrium statistical mechanics (82B20) Statistical mechanics of random media, disordered materials (including liquid crystals and spin glasses) (82D30) Statistical mechanics of magnetic materials (82D40) Numerical algorithms for specific classes of architectures (65Y10)

Cites Work

• Mersenne twister
• Equation of State Calculations by Fast Computing Machines
• Numerical recipes. The art of scientific computing.
• Critical phenomena and renormalization-group theory
• Title not available (Why is that?)
• Title not available (Why is that?)
• A portable high-quality random number generator for lattice field theory simulations
• Tables of linear congruential generators of different sizes and good lattice structure
• GPU accelerated Monte Carlo simulation of the 2D and 3D Ising model
• Multi-GPU accelerated multi-spin Monte Carlo simulations of the 2D Ising model
• Title not available (Why is that?)
• Benchmarking GPU and CPU codes for Heisenberg spin glass over-relaxation
• Title not available (Why is that?)
• Feedback-optimized parallel tempering Monte Carlo
• The critical behavior of 3D Ising spin glass models: universality and scaling corrections
• Parallelizing the cellular Potts model on graphics processing units
• Parallelization of the Ising model and its performance evaluation

Cited In (12)

• Suppressing correlations in massively parallel simulations of lattice models
• Simulation of \(1+1\) dimensional surface growth and lattices gases using GPUs
• Multipath metropolis simulation: an application to the classical Heisenberg model
• Massively parallel multicanonical simulations
• High performance implementations of the 2D Ising model on GPUs
• Benchmarking GPU and CPU codes for Heisenberg spin glass over-relaxation
• IANUS: scientific computing on an FPGA-based architecture
• GPU-accelerated simulation of massive spatial data based on the modified planar rotator model
• Canonical Monte Carlo multispin cluster method
• Massively parallel simulations for disordered systems
• GPU accelerated population annealing algorithm
• Monte-Carlo simulations of Ising-like materials: multipath Metropolis vs cluster update approach

Uses Software

• RANLUX
• CUDA
• MersenneTwister
• OpenCL
• Janus

This page was built for publication: Performance potential for simulating spin models on GPU

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q417832)