STE||AR at Supercomputing 2011

We already wrote about our demos we will show at Supercomputing 2011 next week here. Please come visit us at the LSU booth – 2839. The week at the Supercomputing conference will be an exciting week for our group as it is the first time we present our work to the public. We will demonstrate the HPX (High Performance ParalleX) Runtime System technology which is providing the first freely available, open source implementation of the ParalleX execution model. HPX is solidly based on many years of experience in writing highly parallel applications for HPC systems. It is a modular, feature-complete, and performance oriented representation of the ParalleX execution model targeted at conventional architectures and, currently, Linux based systems, such as SMP nodes and conventional clusters.

ParalleX is an open parallel execution model designed to eliminate the intrinsic constraints of conventional programming methods and more flexibly manage application algorithm parallelism to enable the development of scalable applications for many-core systems. ParalleX represents a paradigm shift from traditional message-passing computing to a message-driven model that focuses on dynamic adaptive locality control and resource management instead of the static data distribution utilized in today’s programming models. It departs from the conventional model of communicating sequential processes by introducing fine-grain parallelism of lightweight threads. This open execution model for parallel computation is designed to enable developers to exploit the full processing power of many-core systems with an unprecedented degree of parallelism.

The development of parallel applications that can effectively scale for optimal execution on many-core systems has presented a daunting challenge in the past. With the current HPC trend of integrating heterogeneous and multicore architectures this challenge has increased multifold.

Our group’s HPX technology provides the first freely available, open source runtime system exhibiting the power of the ParalleX execution model to offer a full system software solution boosting the scalability, performance and efficiency of parallel applications running on many-core architectures.

I strongly believe that our technology has the potential to transform the way we program and run applications today. It will massively increase the possible parallelism and thus the efficiency of our codes. The highly modular structure of HPX guarantees a smooth migration path from today’s systems towards future architectures, which provides a stable implementation platform for application developers over the next years. We are very proud to enable scientists to achieve results today which they couldn’t achieve purely using conventional programming models.

In contrast to today’s standard execution model of communicating sequential processes, the design of HPX serves to expose multiple forms of parallelism, thus hiding latency, reducing starvation and contention in the system. For the first time, HPX is bringing an open source implementation of ParalleX to conventional HPC platforms to help solve the massively parallel, data-intensive problems faced by scientific communities right now.

GD Star Rating
STE||AR at Supercomputing 2011, 5.0 out of 5 based on 1 rating
    This entry was posted in General, SC11 and tagged , by Hartmut Kaiser. Bookmark the permalink.

    About Hartmut Kaiser

    Hartmut is an Adjunct Professor of Computer Science at Louisiana State University. At the same time, he holds the position of a senior scientist at the Center for Computation and Technology (LSU). He received his doctorate from the Technical University of Chemnitz (Germany) in 1988. He is probably best known through his involvement in open source software projects, mainly as the author of several C++ libraries he has contributed to Boost, which are in use by thousands of developers worldwide. His current research is focused on leading the STE||AR group at CCT working on the practical design and implementation of the ParalleX execution model and related programming methods. In addition, he architected and developed the core library modules of SAGA for C++, a Simple API for Grid Applications.

    Leave a Reply

    Your email address will not be published. Required fields are marked *