Wednesday, October 29, 2008

Talk: Thursday, October 30, 2008: High-performance Computing on NVIDIA GPUs


Here's a talk about using graphics cards for non-graphics things. These cards are really amazing quite general purpose parallel computers; you might find it worth your will to check it out to see how these things might serve your own needs.

Lars was once graphics faculty so I'm sure his talk will be quite accessible.



---------- Forwarded message ----------
From: Franc Brglez <>
Date: Wed, Oct 29, 2008 at 12:16 PM
Subject: SEMINAR TOMORROW AFFTERNOON: Thursday, October 30, 2008: High-performance Computing on NVIDIA GPUs
To: Franc Brglez <>
Cc:, "Linda S. Honeycutt" <>,, Carlos Benavente <>


      Department of Computer Science Colloquium Series 2008-2009

                High-performance Computing on NVIDIA GPUs
                         Lars Nyland, nVidia

  Date: Thursday, October 30, 2008
  Time: 02:30 PM (talk begins)
  Place: 3211 EB2, NCSU Centennial Campus

  More info and parking:

    Host: Frank Mueller, Computer Science, NCSU

    NOTE: If you want to schedule a meeting with the speaker,
          please contact the host.
In this talk, I will cover three topics: 1) The NVIDIA GPU computing architecture, 2) The CUDA programming language, and 3) and recent work on N-Body simulation. The GPU architecture supports both graphics and non-graphics computation, using an array of custom processors on a single chip. The programming model is neither SIMD nor MIMD, but somewhere in between, where we can exploit the advantages of each. The current performance part has 240 processors running at 1.5 GHz. With dual-issue capabilities, this places the achieved peak performance just under 1 TFLOP. CUDA is NVIDIA's C/C++ programming language for programming the GPU. It has a few extensions that include thread launch/terminate, synchronization, data sharing, and atomic operations. I'll discuss a collaborative effort with Jan Prins (UNC-CH) and Mark Harris (NVIDIA), where we have written an N-Body simulator using CUDA that runs on NVIDIA GPUs. We achieve a sustained computational rate over 400 GFLOPS. I'll finish with a few demonstration applications, as well as a discussion of how other groups are using NVIDIA GPUs to accelerate their computations. As a postscript, I'll mention the 'professor partnership program' where academicians can receive GPU computing hardware at no cost.

Lars Nyland is a senior architect in the Compute group in NVIDIA's Durham, NC office. He designs, develops and tests architectural features that enable HPC on GPUs. Prior to joining NVIDIA in 2005, he was an associate professor of Computer Science at the Colorado School of Mines. Prior to that, he was a research associate professor at UNC Chapel Hill from 1991 until 2003. He received his Ph.D. from Duke University in 1991 under the supervision of professor John Reif studying parallel programming techniques. In 2000, he worked half-time jumpstarting Deltasphere, Inc., a company that builds scene digitizers (primarily for forensic applications).

Benjamin Watson
Design Graphics Lab
Associate Professor
Dept. Computer Science
North Carolina State University
EBII 2280, 890 Oval Dr, Box 8206
Raleigh, NC 27695-8206
Phone: 919-513-0325
Fax: 919-515-7896
Lab: 919-513-0847

No comments:

Post a Comment