Preliminary design examination of the ParalleX system from a software and hardware perspective

Abstract

Exascale systems, expected to emerge by the end of the next decade, will require the exploitation of billion-way parallelism at multiple hierarchical levels in order to achieve the desired sustained performance. While traditional approaches to performance evaluation involve measurements of existing applications on the available platforms, such a methodology is obviously unsuitable for architectures still at the brainstorming stage. The prediction of the future machine performance is an important factor driving the design of both the execution hardware and software environment. A good way to start assessing the performance is to identify the factors challenging the scalability of parallel applications. We believe the root cause of these challenges is the incoherent coupling between the current enabling technologies, such as Non-Uniform Memory Access of present multicore nodes equipped with optional hardware accelerators and the decades older execution model, i.e., Communicating Sequential Processes (CSP). Supercomputing is in the midst of a much needed phase change and the High-Performance Computing community is slowly realizing the necessity for a new design dogma, as affirmed in the preliminary Exascale studies. In this paper, we present an overview of the ParalleX execution model and its complementary design efforts at the software and hardware levels, while including power draw of the system as the resource of utmost importance. Since the interplay of hardware and software environment is quickly becoming one of the dominant factors in the design of well integrated, energy efficient, large-scale systems, we also explore the implications of the ParalleX model on the organization of parallel computing architectures. We also present scaling and performance results for an adaptive mesh refinement application developed using a ParalleX-compliant runtime system implementation, HPX.

An Autonomic Performance Environment for Exascale

“…The HPX-3 runtime system [3,18,35] serves as a starting point as programming tools and runtime system target at the beginning of the XPRESS project. This has been complemented by the development of HPX-5, which is being developed to add functionality for fault tolerance and power management, and to provide a robust open-source runtime system.…”

Asynchronous Task Scheduling of the Fast Multipole Method Using Various Runtime Systems

“…What has not been studied, however, is the capability and effectiveness of task schedulers when they handle a more sophisticated algorithm like the FMM. This paper attempts to address this issue and considers four different runtime systems, including Cilk [6], [12], C++ thread and future libraries [13], the High Performance ParalleX (HPX-5) library [14], and OpenMP tasks. Here, we use runtime for ease of reference because strictly speaking, C++ thread and future libraries are not a runtime system.…”

PXFS: A persistent storage model for extreme Scale

“…All of these factors (SLOW) are difficult to avoid using today's prevalent programming models, and a new computational strategy is required to achieve dramatic increases in performance. The ParalleX execution model [4]- [6] is offered as a means of addressing these critical computational requirements. It is striving to expose myriad forms of parallelism, hide system wide latencies, decouple hardware execution resources from executing software tasks to prevent the blocking of processor cores, and to enable runtime dynamic adaptive scheduling to employ real-time system state to resource management decisions.…”