TALK ANNOUNCEMENT "Many-core programming challenges and a human-language of programmability" Tim Mattson Intel

11am, Wednesday, April 25th, 2007 Packard Hall 101

ABSTRACT:

Many core processors are here to stay. Four cores per package are available now, and 8 cores are just around the corner. With Moore's Law going strong, we can expect dozens if not hundreds of cores per CPU in the foreseeable future.

The many core hardware challenges are daunting, but they pale next to the software challenges. Many core systems will require high levels of concurrency at every layer of the software infrastructure. Where will we find programmers able to work with so much concurrency?

A nature choice is the HPC community. HPC-programmers have been working with concurrency for decades and are well positioned to play a key role in establishing the many core era. But if they are going to take on this role, HPC programmers must focus on the right problems. They must drop their obsession with performance and focus on the vastly more important issue programmability.

To foster a dialog about programmability, we need a new human-language of programmability; a language that exposes the subjective, cognitive issues in programming systems. Fortunately, a good starting point for this language exists in Thomas Green's cognitive dimensions of programming. If we as a community explore Green's dimensions and modify them to "make them our own", we will have the human-language we need to support a systematic dialog about programmability.

This talk will touch on all of these issues with a clear statement of the software challenges, a realistic prioritization of those challenges, and finally a start at developing our own set of cognitive dimensions for parallel programming. My goal is to start a productive dialog focused on the right problems and leading to collaborations that solve the many core programming problem once and for all.

BIO:

Tim Mattson earned a PhD. for his work on quantum molecular scattering theory (UCSC, 1985). This was followed by a Post-doc at Caltech where he worked on the Caltech/JPL hypercubes. Since then, he has held a number of commercial and academic positions with high performance computers as the common thread. Application areas have included mathematics libraries, exploration geophysics, computational chemistry, molecular biology, and bioinformatics.

Dr. Mattson joined Intel in 1993. Among his many roles at Intel, he was applications manager for the ASCI teraFLOPS project, helped create OpenMP, founded the Open Cluster Group (OSCAR), and launched Intel's programs in computing for the Life Sciences.

Currently, Dr. Mattson is conducting research on abstractions that bridge across parallel system design, parallel programming environments, and application software. This work builds on his recent book on Design Patterns in Parallel Programming (written with Professors Beverly Sanders and Berna Massingill and published by Addison Wesley). The patterns provide the "human angle" and help keep his research focused on technologies that help general programmers solve real problems.