Site Archive (Complete)
DrDobbs Portal Blog: Parallelism: It's Everywhere, It's Everywhere
EDITOR'S EYE

The World of Software Development.

by Jon Erickson
April 30, 2008

Parallelism: It's Everywhere, It's Everywhere

Parallel computing is everywhere these days. You might even say that it's "pervasive." Hmmm...pervasive parallelism. Has a nice ring, eh? Well, apparently the folks at Stanford University thought enough of the alliteration that they used it to name the Pervasive Parallelism Lab (PPL).

The goal of the Lab, which is led by Professor Kunle A. Olukotun of the Stanford Electrical Engineering and Computer Science department, is two-fold: Make parallel programming accessible to the average programmer, and develop algorithms, environments, and runtime systems that scale to thousands of hardware threads. To bridge the gap between where parallelism is today and where PPL hopes it might be a few years from now, PPL team members are focusing on: Education (moving parallelism into a mainstream computer science curriculum), programming paradigms (domain-specific languages and high-level abstractions), architectures (hardware support for new paradigms), and applications (in business, games, embedded, and the like).

Joining Stanford in this venture are companies such as NVIDIA, AMD, Hewlett Packard, IBM, Intel, and Sun Microsystems. While the involvement of processor companies like Intel and AMD rings familiar, the role of companies such as NVIDIA is less so. But NVIDIA has gone a long way in solving computationally intensive problems with its GPUs and software, specifically the CUDA programming environment that gives developers access to the GPU's massively parallel architecture via a familiar C-language environment. (It's worth noting that Dr. Dobb's recently launched the article series "CUDA, Supercomputing for the Masses"; see Part I here and Part II here.)

So is all the buzz about parallelism the real deal or is it just more PR? Well, Bill Dally, chair of the computer science department at Stanford, thinks it's the real deal: "Parallel programming is perhaps the largest problem in computer science today and is the major obstacle to the continued scaling of computing performance that has fueled the computing industry, and several related industries, for the last 40 years." Sounds real to me, too.

-- Jonathan Erickson
jerickson@ddj.com

Posted by Jon Erickson at 03:35 PM  Permalink




 

♦ sponsored
INFO-LINK


Related Sites: DotNetJunkies, SD Expo, SqlJunkies