Special Issue: Manycore and Accelerator-based High-performance Scientific Computing Introduction

In January 2011, the International Center for Computational Science (ICCS) at LBNL and UC Berkeley organized a workshop in Berkeley, CA, USA on the role of emerging many-core and accelerator-based architectures in science and technology. The focus of the workshop was to usher experts and enthusiasts; from disparate backgrounds in academia and industry to introduce, explore and discuss the scope and challenges of harnessing the full potential of these novel architectures for high-performance computing (HPC), especially in the broad range of scientific and technology disciplines. The emerging new techniques open a route for studying future generations of hardware and software using silicon and electrical power much more efficiently, a necessity on the path to exascale computing. Emerging many-core and accelerator-based architectures offer unprecedented potential for energy-efficient HPC in science and technology. Limited early adoption of such architectures by a broad range of compute- and data-intensive applications has distinctly demonstrated unmatched gains in acceleration and power consumptions. This workshop strived to set the stage for academic and industry leaders to enable partnerships for growth of research and innovation in the broader community. Early adopting core-science teams from across the world had an opportunity to get together to coordinate next steps and establish a regular transcontinental forum in the coming years. This special issue of JHPCA displays a collection of peer-reviewed articles presented at the ICCS workshop that demonstrate the advanced use of the new architectures and associated programming models for a wide range of applications. The goal of the special issue of the journal is to demonstrate the challenges and solutions in adopting the technologies common to a broader applications base.