A lightweight library for building scalable tools

Authors:
Emily R. Jacobson;Michael J. Brim;Barton P. Miller
Affiliations:
Computer Sciences Department, University of Wisconsin, Madison, Wisconsin;Computer Sciences Department, University of Wisconsin, Madison, Wisconsin;Computer Sciences Department, University of Wisconsin, Madison, Wisconsin
Venue:
PARA'10 Proceedings of the 10th international conference on Applied Parallel and Scientific Computing - Volume 2
Year:
2010

Citing 4
Cited 0

MRNet: A Software-Based Multicast/Reduction Network for Scalable Tools

Proceedings of the 2003 ACM/IEEE conference on Supercomputing
The Tau Parallel Performance System

International Journal of High Performance Computing Applications
Lessons learned at 208K: towards debugging millions of cores

Proceedings of the 2008 ACM/IEEE conference on Supercomputing
Overcoming Scalability Challenges for Tool Daemon Launching

ICPP '08 Proceedings of the 2008 37th International Conference on Parallel Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

MRNet is a software-based multicast reduction network for building scalable tools. Tools face communication and computation issues when used on large systems; MRNet alleviates these issues by providing multicast communication and data aggregation functionalities. Until now, the MRNet API has been entirely in C++. We present a new, lightweight library that provides a C interface for MRNet back-ends, making MRNet accessible to a wide range of new tools. Further, this library is single threaded to accommodate even more platforms and tools where this is a limitation.This new library provides the same abstractions as the C++ library, using an API that can be derived by applying a standard translation template to the C++ API.