"This year's Annual

Meeting will give

attendees...a high-

quality Technical Program

with more posters."

-Annual Meeting General Chairman, Ken Helm

 

 

W-14 GPU Computing for Upstream Applications and HPC User Tools

Date: Friday, 23 September
Time: 8:30 a.m.-5 p.m.
Henry B. Gonzalez Convention Center, Room 214 D
Organizer: The Society of HPC Professionals (SHPCP)

Description:

This full day workshop is being jointly organized by The Society of HPC Professionals and the HPC Advisory Council and will be conducted in the following two sessions:

Session 1: GPU Computing for Upstream Applications (Morning)
The rapid increase in the performance of graphics hardware, coupled with recent improvements in its programmability, has made graphics accelerators a compelling platform for computationally demanding tasks in a wide variety of application domains, including upstream applications. The modern GPU is a highly data-parallel processor, optimized to provide very high floating-point arithmetic throughput for problems suitable to solve with a single program multiple data model.

GPU clusters present new challenges – how to build a balanced hybrid system, how to create an optimized GPU to CPU communications, how to balance the compute nodes/GPU ratio and more. The workshop will cover concepts and topologies for hybrid clusters of GPU/CPU, the potential benefits and best practices for building and using such systems. The future of GPU computing and potential accelerations for GPU communications models will also be covered.

Session 2: HPC User Tools (Afternoon)
Petascale computing is coming to the upstream industry; and along with such computational advancement comes the scale of algorithmic and application development and debugging challenges never seen before nor experienced. With processor counts growing rapidly and GPUs emergence comprising very large hybrid systems, what happens when the software fails to perform adequately or simply does not work? Couple this with the waning of technical experts available for performance tuning and optimization across a wide variety of hardware environments make the criticality of useful and effective HPC tools paramount.

This workshop assesses the effectiveness of tools and techniques that are or could be helpful in upstream Petascale deployments and outlines potentially beneficial tools areas for research and development. This workshop is designed to foster a network of experts who are able to share information as Petascale systems come on line. The technology future of the HPC user tools and its marketplace will be covered.

Back to Technical Programs Workshops menu

Register Today!