Computer Science and Mathematics Projects
The XPRESS Project
Exascale systems, with a thousand times the compute capacity of today’s leading edge petascale computers, are expected to emerge during the next decade. Their software systems will need to facilitate the exploitation of exceptional amounts of concurrency in applications, and ensure that jobs continue to run despite the occurrence of system failures and other kinds of hard and soft errors. Adapting computations to changes in the available resources is likely to become the norm. As a result, considerable innovation is required to develop system support to meet the needs of future computing platforms.
The XPRESS project aims to develop and prototype a revolutionary software system for extreme-scale computing for both exascale and strong-scaled problems. XPRESS is exploring a set of innovations in execution models, programming models and methods, runtime and operating system software, dynamic and adaptive scheduling and resource management algorithms and mechanisms, and instrumentation and introspection techniques to achieve unprecedented efficiency, scalability, and programmability in the context of billion-way parallelism while providing seamless migration of legacy application codes. The strong vertically integrated research team brings complementary capabilities and accomplishments to achieve comprehensive coverage of this complex research and design space.
Translating OpenMP to OpenX Stack
XPRESS intends to not only carry out research to determine how best to enable exascale performance capability for DOE applications, both current and future, but also to develop and deliver a practical computing system software X-stack, OpenX, for future practical DOE exascale computing systems. It will moreover provide programming methods and environments for effective means of expressing application and system software for portable exascale system execution.
The Brookhaven team contributes to the XPRESS effort by performing work that will evaluate the ability of legacy applications to execute in an exascale software environment. With a focus on applications that employ OpenMP to specify intra-node parallelism, they are developing an implementation of OpenMP that targets OpenX and are evaluating its usage. Together with project partners, they are creating applications that use features of OpenMP that have been recently introduced into the specification and that may be important for exascale applications. They are also actively evaluating potential alternative approaches to providing system support at this scale, with a view to understanding the implications for legacy codes based upon MPI and OpenMP. Based upon these experiences, they are considering potential extensions of OpenMP that might facilitate full exploitation of the capabilities of exascale systems by legacy codes with minimal adaptation.
XPRESS was funded through the DOE Office of Science Office of Advanced Scientific Computing Research via the 2012 X-Stack Program of Programming Challenges, Runtime Systems, and Tools.
