June 16-20, 2013

Leipzig, Germany

Presentation Details

Name: Fork-Join & Data-Driven Execution Models on Multi-Core Architectures: Case Study of the FMM
Time: Thursday, June 20, 2013
9:40 AM - 10:00 AM
Room:   Hall 5
CCL - Congress Center Leipzig
Speakers:   Amer Abdelhalim, Tokyo Institute of Technology
Abstract:   We present a study of fork-join and data-driven execution models on multi-core architectures at the level of task parallelism. For this purpose, we use a highly optimized fork-join based implementation of the FMM and extend it to a data-driven implementation using a distributed task scheduling approach. We expose some limitations of the conventional fork-join implementation in terms of synchronization overheads. We find that these are not negligible and their elimination by the data-driven method, with a careful data locality strategy, was beneficial. Experimental evaluation of both methods on state-of-the-art multi-socket multi-core architectures showed up to 22% speed-ups of the data-driven approach compared to the original method. We demonstrate that a data-driven execution of FMM not only improves performance by avoiding global synchronization overheads but also reduces the memory-bandwidth pressure caused by memory-intensive computations.

Paper authors:
Abdelhalim Amer1, Naoya Maruyama2, Miquel Pericas1, Kenjiro Taura3, Rio Yokota4, Satoshi Matsuoka1
1Tokyo Institute of Technology, Tokyo, Japan
2RIKEN, Kobe, Japan
3The University of Tokyo, Tokyo, Japan
4KAUST, Saudi Arabia
Program may be subject to changes.