Optimizing the advanced accelerator simulation framework synergia using OpenMP

Authors:
Hongzhang Shan;Erich Strohmaier;James Amundson;Eric G. Stern
Affiliations:
Future Technology Group, Computational Research Division, Lawrence Berkeley National Laboratory, Berkeley, CA;Future Technology Group, Computational Research Division, Lawrence Berkeley National Laboratory, Berkeley, CA;Fermi National Accelerator Laboratory, Batavia, IL;Fermi National Accelerator Laboratory, Batavia, IL
Venue:
IWOMP'12 Proceedings of the 8th international conference on OpenMP in a Heterogeneous World
Year:
2012

Citing 7
Cited 0

Synergia: An accelerator modeling tool with 3-D space charge

Journal of Computational Physics
Dynamic Task and Data Placement over NUMA Architectures: An OpenMP Runtime Perspective

IWOMP '09 Proceedings of the 5th International Workshop on OpenMP: Evolving OpenMP in an Age of Extreme Parallelism
Three-level hybrid vs. flat MPI on the Earth Simulator: Parallel iterative solvers for finite-element method

Applied Numerical Mathematics - 6th IMACS International symposium on iterative methods in scientific computing
Performance analysis of large-scale OpenMP and hybrid MPI/OpenMP applications with VampirNG

IWOMP'05/IWOMP'06 Proceedings of the 2005 and 2006 international conference on OpenMP shared memory parallel programming
A programming model performance study using the NAS parallel benchmarks

Scientific Programming - Exploring Languages for Expressing Medium to Massive On-Chip Parallelism
Hybrid programming model for implicit PDE simulations on multicore architectures

IWOMP'11 Proceedings of the 7th international conference on OpenMP in the Petascale era
Critical path-based thread placement for NUMA systems

Proceedings of the second international workshop on Performance modeling, benchmarking and simulation of high performance computing systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

Synergia is an advanced accelerator simulation framework widely used in the accelerator community. Unfortunately, its performance and scalability suffers significantly from very high communication requirements. In this paper, we address this issue by replacing the flat MPI programming model with the hybrid OpenMP+MPI programming model. We describe in detail how the code has been parallelized in OpenMP and what the challenges are. The improved hybrid code can perform over 1.7 times better than the original program for a realistic benchmark problem.