About

Hardware accelerators have the potential to elevate scientific computing to new levels of performance. This one-day symposium will explore the use of GPUs, CELL processors, FPGAs and multi-core CPUs for large-scale scientific computing. SHARCNET will soon deploy high-performance clusters containing both CELL and GPU accelerators, and this symposium will give SHARCNET researchers the chance to learn about these new technologies from keynote speakers who are at the forefront of research in this field. SHARCNET researchers who have first experiences in this new area are invited to contribute presentations and posters.

Keynote Speakers:

• Ben Bergen, Los Alamos
  
  • Presentation: Heterogeneous Petascale computing with RoadRunner
  • Further Information: RoadRunner
• Michael McCool, University of Waterloo
  
  • Presentation:The RapidMind Platform for Portable Programming of Multi-core Processor and Many-core Accelerators
    
  • Further Information: RapidMind

Vendor Presentations:

• Jonathan Cohen, NVIDIA
  
  • Presentation: Parallel Computing with CUDA
  • Further Information: CUDA
• Michael Perrone, IBM
  
  • Presentation: Feeding the Multicore Beast
    
  • Further Information: Cell Broadband Engine
• Jason Yang, AMD
  
  • Presentation: AMD Stream Computing
  • Further Information: AMD Stream Computing

Contributed Presentations:

• Christopher Kumar Anand
  Coconut Multi-Core Run Time for Cell BE
• Gabriel Esteves
  An implementation of the quasi-Monte Carlo Baum-Welch algorithm using the stream processing model
• Gabriel Grant
  SUITCASE - SPU Utility for Image Test Creation and Scaffold Erecting
• Michael Kinsner
  CUDA-based Acceleration of a Scale-Space Image Feature Detector
• Chris Madill
  An MPI Approach to High-Performance Computing with FPGAs
• Hugh Merz
  Overview of accelerator technologies and SHARCNET resources
• Alejandro Salinger
  Optimal Speedup on a Low-Degree Multi-Core Parallel Architecture (LoPRAM)
• Meilian Xu
  Cell Processing for two Scientific Kernels

Contributed Posters:

• Marina L Suominen Fuller
  "Science Studio": Remote control of .big science. facilities and data processing/data storage requirements
• Stephen M. J. Guzik
  Computational Enhancement of High-Order Residual Distribution Schemes Using Graphics Processing Units
• Jeff Kingyens
  General Purpose GPU (GPGPU) Programming Models for FPGAs
• Zsolt Zsoldos
  eHiTS Lightning on the Cell processor

Important Dates:
April 28 : deadline for presentation abstract submission
May 9 : final program announcement
May 15 (midnight): deadline for registration and poster submissions (registration is free but required)
May 27 : SHARCNET Symposium on GPU and CELL Computing

Organizing Committee:
Hans De Sterck, University of Waterloo
Ilias Kotsireas, Wilfrid Laurier University
Hugh Merz, SHARCNET

Financial sponsors:

• Dean of Mathematics, University of Waterloo
• Dean of Science, Wilfrid Laurier University
• Research Office, Wilfrid Laurier University
• SHARCNET

Facility sponsors:

• University of Waterloo
• Institute for Computer Research, University of Waterloo