Parallel Programming Talk - Dr. David Patterson on the future of Parallel Programming

By Aaron Tersteeg (Intel) (151 posts) on June 10, 2009 at 3:12 pm

Aaron & Clay talked with Dr. David Patterson on the 34th episode of Parallel Programming Talk.

Download the MP3 of the show.

Download link to a high quality MP4 video file of the show (27min, 314MB)

News:

Intel Software Network Television (ISNTV) officially launched on June 9th. The day's live programming started off with Parallel Programming Talk followed at 10:00AM(PST) with Teach Parallel, 2:30PM(PST) with Visualize This and a 4:00PM(PST) special show on Moblin. Follow the links and check out instant replays of all the shows.

Threading Challenge Update

June 15 – Threading Challenge #6 starts

June 19 – Threading Challenge #5 due

July 3 - Threading Challenge #6 due

Three contest have been judged. Congratulations to "denghui0815" who is our 1st & 3rd problem winner and "haojn" who won our second problem. It's not to late to get in the challenge and win individual challenge prizes.

Listener Question Show

Our next listener question show is July 7 – send in questions or idea for our show to parallelprogrammingtalk@intel.com.

Upcoming Events

Clay’s book is now shipping

The Art of Concurrency

A Thread Monkey's Guide to Writing Parallel Applications

By Clay Breshears

On Today's Show

Dr. David Patterson is the Director of the Universal Parallel Computing Research Center at UC Berkeley sponsored by Intel and Microsoft.

He is also the UC Berkeley Pardee Chair of Computer Science and pioneering expert in computer architecture(RISC, RAID, NOW). He’s served on the Information Technology Advisory Committee to the U.S. President, Chair of the Computer Research Association, and as President of ACM.

Here are a few of the questions and areas discussed:

David it sounds like you’ve worked on many amazing projects. Where does your passion lie these days?

It has been a little over a year since the UPCRC was launched. How has it been going? What do you consider your key accomplishments? Unexpected learning’s?

Since 2005 you have been working on changing the conventional wisdom of parallel processing.

You framed the landscape issues with 7 areas of focus (the list has grown) for the design and evaluation of parallel programming models and architectures (the dwarfs). Could you give us an example of a one of the dwarfs and how they are being addressed?

Your Research agenda has many components that can be broken down into Layers (from Application to Architecture) and bands (such as Diagnosing and Correctness).  What are some of the applications?

I heard that Berkeley did a UPCRC review last week at your Par Lab retreat. What was covered? How did it go?

What are the challenges of future Parallel Architecture?

  • Manycore, not multi-core
  • Autotuners, not compliers
  • Human-centric, not machine-centric programming models

What do you see as the programming model of the future?

Later this summer I heard that Berkeley is hosting a Parallel Programming Laboratory Boot Camp from August 19-21. What is the format, who should attend?

Where can our listeners go to find out more about the Universal Parallel Computing Research Center at UC Berkeley?

Suggested Reading:

The Parallel Computing Laboratory at U.C. Berkeley: A Research Agenda

Based on the Berkeley View (EECS-2008-23)

By Krste Asanovic, Ras Bodik, James Demmel, Tony Keaveny, Kurt Keutzer, John D. Kubiatowicz, Edward A. Lee, Nelson Morgan, George Necula, David A. Patterson, Koushik Sen, John Wawrzynek, David Wessel and Katherine A. Yelick

Roofline: an insightful visual performance model for multicore architectures

The Roofline model offers insight on how to improve the performance of software and hardware. The paper demonstrates this model on four diverse multicore computers using four key floating-point kernels.

By Samuel Williams, Andrew Waterman & David Patterson

Communications of the ACM archive

Volume 52 ,  Issue 4  (April 2009)

Coming Up Next on Parallel Programming Talk:

  • June 16 – Duncan McCallum (cilk.com)
  • June 23 - Calvin Lin (University of Texas)
  • June 30 – Joe Duffy (Microsoft)
Categories: Parallel Programming

For more complete information about compiler optimizations, see our Optimization Notice.

Comments (0)

Trackbacks (6)


Leave a comment  

To obtain technical support, please go to Software Support.
Name (required)*

Email (required; will not be displayed on this page)*

Your URL (optional)


Comment*