Productive parallel programming sounds like an oxymoron to most of us, especially those of us who've learned to write parallel programs in the pre-multi/many-core age. The practices of productive programming have diverged from practices that allow tools (like compilers) can effectively take advantage of performance features of processors, like multi-core and vector parallelism. These practices appear hopelessly irreconcilable, but a return to dynamic generative programming methods can help (while simultaneously overturning the conventional wisdom that dynamic compilation and managed runtimes equate to low performance). Intel's Ct technology, which is currently being productized, aims to both provide a tool for developers to write parallel programmers productively and create an infrastructure for implementation of other data parallel domain-specific (aka productivity) libraries and languages. In this talk, I will describe the Ct technology, it's goals, implementation, and the possibilities opened up by this approach.
There is no downloadable version of the slides for this talk available at this time.
About the speaker:
|Anwar is a Principal Engineer and Engineering Manager with Intel's Software and Services Group, architecting and developing a new data parallel programming tool based on Intel's Ct Technology (which he developed while in Intel Labs). In the labs, he has worked on diverse topics such as parallel language and compiler design, parallel architecture evaluation, optimizing memory system performance, and multimedia applications. Anwar Ghuloum earned degrees at the University of California, Los Angeles (B.S., Computer Science and Engineering) and Carnegie Mellon University's School of Computer Science (Ph.D., Computer Science, 1996), where his thesis introduced concepts of Nested Data Parallel idioms to traditional automatic parallelizing compilers. Before joining Intel, he co-founded and was the CTO of a fab-less semiconductor startup called Intensys that built programmable, highly parallel image and video processors for the consumer electronics market. Prior to that, Anwar developed novel predictive drug design software for early lead optimization using 3D surface pattern recognition techniques for a biotech startup called MetaXen (acquired by Exelexis Pharmaceuticals). He has also served as a post-doctoral research associate at Stanford University's Computer Science department. A recurring theme in Anwar's work has been to bridge high-level application knowledge and low-level parallel architecture constraints with careful parallel language and compiler design to achieve the optimal trade-offs in productivity and performance.|
Anwar M Ghuloum