TUTORIALS

Tutorial 1: Success Secrets of Managing Enterprise Architecture Projects
Tutorial 2: Using the Pilot Library: A Fresh Alternative to MPI for HPC Clusters
Tutorial 3: An Introduction to CUDA: Harnessing Graphics Processing Units for High-Throughput General-Purpose Computation

Alexei Botchkarev, Chair, IEEE Toronto Section Wishes IEEE A Happy 125th Anniversary and Invites to the TIC-STH 2009
(Sorry for the 2008 IEEE Sections Congress lobby noise in the background)

Benefits of Attending

Prestigious International IEEE Conference
Full-Paper Peer Review
Publication in the IEEE Xplore – Most Respected Scientific Full-Text Database
Attend the Conference On-Site in Toronto or Participate On-Line
Low Registration Fee – $299
Opportunity to Publish Lengthier Papers – Up to 6 (six) Pages With No Extra Fee
High-Value Tutorials Included with Registration
Conference is focusing on advanced interdisciplinary problems across a broad spectrum of the IEEE fields of interest. The scope is NOT limited to the traditional IEEE areas – electrical, electronic, computing, engineering.
And more...

TUTORIAL 3:

Title: An Introduction to CUDA: Harnessing Graphics Processing Units for High-Throughput General-Purpose Computation

Speaker: Lukasz Wawrzyniak
University of Guelph, Ontario

» Register here
Deadline: September 21, 2009, 11:59PM EDT

Date: Sunday, September 27, 2009
Time: 9:30 AM; Duration – 3 hours
Attend onsite: 245 Church St. (Ryerson University), Toronto
Attend online: Web conferencing option available

Cost

Conference Attendees: Free (requires registration ID number)
IEEE and/or PMI members: $55 (Canadian)
All others: $75 (Canadian)

Speaker

Lukasz Wawrzyniak obtained his BSc in Applied Computer Science from Ryerson University. He obtained his MSc in Computing and Information Science from the University of Guelph, where he is currently a PhD candidate. His research focuses on detecting significant high-density clusters in multidimensional data with applications to disease surveillance. One important avenue of this research concerns accelerating algorithms using graphics processing units and a variety of computing platforms including CUDA, ATI Stream, and OpenGL.

Course Overview

CUDA is a computing platform that exposes NVIDIA graphics cards as general-purpose compute devices. The computational capabilities of today's massively parallel graphics processing units (GPUs) can be harnessed to accelerate data-parallel algorithms. CUDA has been used successfully by researchers in a wide variety of disciplines including biochemistry, astrophysics, geology, visualization, and countless others. This tutorial is an introduction to the CUDA programming model. It examines a typical "Hello, World" application and introduces some of the more advanced features using increasingly sophisticated examples. The goal is to give the attendee a broad perspective on computing with CUDA as well as sufficient information to start developing simple CUDA applications.

What You Will Learn

Brief history of GPGPU (general-purpose computing on graphics processing units)
Conceptual overview of CUDA and its relation to other GPGPU approaches
The stream processing model
Using CUDA shared memory for inter-thread communication
Optimizing memory access patterns to accelerate applications
Evaluating performance
General optimization strategies

Prerequisites

This is an introductory tutorial and familiarity with any form of GPU programming is not required. However, the code examples assume exposure to C programming.