BEGIN:VCALENDAR
PRODID:-//Microsoft Corporation//Outlook MIMEDIR//EN
VERSION:1.0
BEGIN:VEVENT
DTSTART:20111116T193000Z
DTEND:20111116T200000Z
LOCATION:TCC 303
DESCRIPTION;ENCODING=QUOTED-PRINTABLE:ABSTRACT: The growing need for power efficient extreme-scale high-performance computing coupled with plateauing clock-speeds is driving the emergence of massively parallel compute architectures. For example, many-core graphics processing units (GPUs) offer thousands of SIMD cores primarily designed for image and video rendering are used for high-performance scientific computations. The new architectures typically offer ANSI standard programming models such as CUDA (NVIDIA) and OpenCL. However, the wide-ranging adoption of the parallel architectures requires reengineering of existing libraries leading to expensive and error-prone code rewrites with no guarantee of speedups.=0A =0AWe use GPU-aware AMR based common-codebase, GAMER, to examine the unique approach of solving multi-science problems in astrophysics, hydrodynamics and particle physics. We demonstrate significant speedups in broad and disparate class of scientific applications. By extensively reusing the extendable single common-codebase we mitigate the impediments of significant code rewrites on NVIDIA C2050 GPU and Intel 8-core Nehalem CPU on Dirac cluster at NERSC.
SUMMARY:Multi-Science Applications with Single Codebase - GAMER - for Massively Parallel Architectures
PRIORITY:3
END:VEVENT
END:VCALENDAR