Early experiments with the OpenMP/MPI hybrid programming model

Authors:
Ewing Lusk;Anthony Chan
Affiliations:
Mathematics and Computer Science Division, Argonne National Laboratory;ASCI FLASH Center, University of Chicago
Venue:
IWOMP'08 Proceedings of the 4th international conference on OpenMP in a new era of parallelism
Year:
2008

Citing 5
Cited 12

Enabling technologies for petaflops computing

Enabling technologies for petaflops computing
MPI versus MPI+OpenMP on IBM SP for the NAS benchmarks

Proceedings of the 2000 ACM/IEEE conference on Supercomputing
A Trace-Scaling Agent for Parallel Application Tracing

ICTAI '02 Proceedings of the 14th IEEE International Conference on Tools with Artificial Intelligence
Toward Scalable Performance Visualization with Jumpshot

International Journal of High Performance Computing Applications
An efficient format for nearly constant-time access to arbitrary time intervals in large trace files

Scientific Programming - Large-Scale Programming Tools and Environments

Hybrid Message-Passing and Shared-Memory Programming in a Molecular Dynamics Application On Multicore Clusters

International Journal of High Performance Computing Applications
Case study for running HPC applications in public clouds

Proceedings of the 19th ACM International Symposium on High Performance Distributed Computing
Hybrid programming model for implicit PDE simulations on multicore architectures

IWOMP'11 Proceedings of the 7th international conference on OpenMP in the Petascale era
Cache size in a cost model for heterogeneous skeletons

Proceedings of the fifth international workshop on High-level parallel programming and applications
Enabling low-overhead hybrid MPI/OpenMP parallelism with MPC

IWOMP'10 Proceedings of the 6th international conference on Beyond Loop Level Parallelism in OpenMP: accelerators, Tasking and more
Portable explicit threading and concurrent programming for MPI applications

PPAM'11 Proceedings of the 9th international conference on Parallel Processing and Applied Mathematics - Volume Part II
Coalition threading: combining traditional andnon-traditional parallelism to maximize scalability

Proceedings of the 21st international conference on Parallel architectures and compilation techniques
An evaluation of user-level failure mitigation support in MPI

EuroMPI'12 Proceedings of the 19th European conference on Recent Advances in the Message Passing Interface
Concurrent programming constructs for parallel MPI applications

The Journal of Supercomputing
Multi-level parallelism for incompressible flow computations on GPU clusters

Parallel Computing
An evaluation of User-Level Failure Mitigation support in MPI

Computing
Performance metrics in a hybrid MPI-OpenMP based molecular dynamics simulation with short-range interactions

Journal of Parallel and Distributed Computing

Quantified Score

Hi-index	0.00

Visualization

Abstract

The paper describes some very early experiments on new architecturesthat support the hybrid programming model. Our results arepromising in that OpenMP threads interact with MPI as desired, allowingOpenMP-agnostic tools to be used. We explore three environments:a "typical" Linux cluster, a new large-scale machine from SiCortex, andthe new IBM BG/P, which have quite different compilers and runtimesystems for both OpenMP and MPI. We look at a few simple, diagnosticprograms, and one "application-like" test program. We demonstrate theuse of a tool that can examine the detailed sequence of events in a hybridprogram and illustrate that a hybrid computation might not alwaysproceed as expected.