Hybrid MPI/StarSs - a case study
Jose Gracia, Christoph Niethammer, Manuel Hasert, Steffen Brinkmann,, Rainer Keller, Colin W. Glass
TL;DR
This paper explores the integration of MPI with the StarSs task parallel model in high-performance computing, demonstrating its application to a Lattice-Boltzmann code and comparing its performance with MPI/OpenMP and MPI-only implementations.
Contribution
It introduces a hybrid MPI/StarSs programming model, discusses its benefits and challenges, and provides a performance comparison with existing hybrid and MPI-only approaches.
Findings
MPI/StarSs hybrid shows promising performance benefits.
Challenges include porting complexity and data dependency management.
Performance comparison highlights strengths and limitations of each approach.
Abstract
Hybrid parallel programming models combining distributed and shared memory paradigms are well established in high-performance computing. The classical prototype of hybrid programming in HPC is MPI/OpenMP, but many other combinations are being investigated. Recently, the data-dependency driven, task parallel model for shared memory parallelisation named StarSs has been suggested for usage in combination with MPI. In this paper we apply hybrid MPI/StarSs to a Lattice-Boltzmann code. In particular, we present the hybrid programming model, the benefits we expect, the challenges in porting, and finally a comparison of the performance of MPI/StarSs hybrid, MPI/OpenMP hybrid and the original MPI-only versions of the same code.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsParallel Computing and Optimization Techniques · Advanced Data Storage Technologies · Lattice Boltzmann Simulation Studies