Monday November 24th Nvidia Seminar - Please attend and pass on

[Connors]

DATE: Monday November 24th, 11:00am

LOCATION: 118 Shepardson


TITLE:

CUDA: Accelerating Science With Massively Parallel Computing

ABSTRACT:

Modern GPUs (graphics processing units) provide a level of massively
parallel computation that was once the preserve of supercomputers like
the MasPar and Connection Machine. NVIDIA's Tesla architecture GPUs
are a fully programmable, massively multithreaded chip with up to 240
cores, 30,720 threads and capable of performing up to a trillion
operations per second. The CUDA platform provides a scalable parallel
programming model consisting of minimal but expressive changes to the
familiar C/C++ language, allowing programmers to focus on parallel
algorithm rather than the mechanics of a new programming language.
Using this platform, researchers across science and engineering are
accelerating applications in their discipline by up to two orders of
magnitude.

In this talk, I will motivate GPU computing and explore the transition
it represents in massively parallel computing: from the domain of
supercomputers to that of commodity "manycore" hardware available to
all. I will also survey some of the results researchers have achieved
in actual science and engineering codes. Finally, I will discuss the
goals, implications, and key abstractions of the CUDA programming
model, and close with some comments on future directions for GPU
computing.

SPEAKER:

Dr. David Luebke
Manager, NVIDIA Research
NVIDIA Corporation
http://luebke.us

David Luebke joined NVIDIA Corporation in 2006 to help found NVIDIA
Research. Previously he spent eight years on the faculty of the
University of Virginia. He received his Ph.D. in Computer Science from
the University of North Carolina in 1998 under the supervision of Dr.
Frederick P. Brooks, Jr. Luebke's principal research interests are
general-purpose GPU computing and realistic real-time computer
graphics. At the University of Virginia he received both the National
Science Foundation CAREER award and the Department of Energy Early
Career Principal Investigator award, as well as the Test of Time award
ACM Symposium on Interactive 3D Graphics "Test of Time Award" in 2005
for the paper with the most impact from the early years of the
Symposium. Together with his colleagues Dr. Luebke has worked on
dozens of papers, articles, chapters, and patents; a short film in the
SIGGRAPH 2007 Electronic Theater; the book "Level of Detail for 3D
Graphics"; and the 2003 Virtual Monticello exhibit seen by over
110,000 visitors to the New Orleans Museum of Art.


Click on http://groups.google.com/group/csu-fall2008-multicore-gpu-class/web/schedule
- or copy & paste it into your browser's address bar if that doesn't
work.

[Connors]

SPEAKER: Dr. David Luebke (Hosted by Dan Connors -
dcon...@colostate.edu)

DATE: Monday November 24th, 11:00am

LOCATION: 118 Shepardson, Colorado State University
(http://www.map.colostate.edu/
)

Click on http://groups.google.com/group/csu-fall2008-multicore-gpu-class/web/schedule?hl=en

[Connors]

Class on December 1st will be the ISTEC seminar (so no 9am meeting):

“Making Parallel and Distributed Computing Systems Robust" by
Professor H. J. Siegel

December 1, 2008, 11:00 – 11:50 a.m.
118 Shepardson

Click on http://groups.google.com/group/csu-fall2008-multicore-gpu-class/web/schedule

[Connors]

Multicore Programming Models and their Implementation Challenges


Monday
Dec 8, 2008
11 - 11:50 AM

Lory Student Center 203



Click on http://groups.google.com/group/csu-fall2008-multicore-gpu-class/web/schedule?hl=en

[Connors]

Posted new schedule items:

Wednesday Dec 10th, 9am (IN CLASS) last day presentation by Gautum
Gupta.

Monday Dec 15th 11am, AMD talk:
Abstact:

The rapid advances in the visual compute capabilities of the commodity
PC platform are bringing forth a new era of Heterogeneous computing.
The
combination of powerful multi-core CPUs and multi-teraflop GPUs
provide
a significant acceleration opportunity for many workloads. Innovative
parallel programming models are also emerging to be the main focus in
the software community. This talk compares and contrasts the GPU with
CPU architectures and also introduces some of the programming
interfaces
such as Brook+/CAL, Dx11 Compute Shaders and OpenCL for GPU
architectures.

Click on http://groups.google.com/group/csu-fall2008-multicore-gpu-class/web/schedule

[Connors]

Speaker: Michael Mantor
Senior GPU Compute Architect / Fellow
AMD Graphics Product Group

Date: Monday, December 15th
Time: 11:00am Reception, 11:30am Seminar

Location: Lory Student Center, RM 213-215
http://www.map.colostate.edu/map-right.aspx#68

Title: Entering the Golden Age of Heterogeneous Computing. (Multi-core
CPUs and Multi-teraflops GPUs, & AMD Stream Computing)

Abstact:
The rapid advances in the visual compute capabilities of the commodity
PC platform are bringing forth a new era of Heterogeneous computing.
The
combination of powerful multi-core CPUs and multi-teraflop GPUs
provide
a significant acceleration opportunity for many workloads. Innovative
parallel programming models are also emerging to be the main focus in
the software community. This talk compares and contrasts the GPU with
CPU architectures and also introduces some of the programming
interfaces
such as Brook+/CAL, Dx11 Compute Shaders and OpenCL for GPU
architectures.

Bio:
Mike has been developing algorithms in software and hardware for
graphics and imaging systems for 20+ years. Starting with the
development of image processing, tracking and recognition systems for
Lockheed Martin, to high end image generators for GE. Mike began
commercial development of commodity graphics for chips such as Intel
740, all ATI/AMD RadeonTM Series graphics devices and Xbox 360 console
graphics. Mike has been a key contributor in texture subsystems,
rasterizers, ATI's first programmable vertex shaders and ATI/AMD's
Unified Programmable Shaders systems with a heavy emphasis in the
development of the ALU units. Mike has accumulated 20+ issued patents,
10+ pending patents and has spoken at several leading universities
such
as MIT, Stanford, and UCF and provided technical talks at Graphics
Hardware and Hot Chips.

[Connors]

Hopefully the snow doesn't delay things.... it should be a VERY good
talk tomorrow. Please tell your colleagues to come.

Click on http://groups.google.com/group/csu-fall2008-multicore-gpu-class/web/schedule?hl=en