| advanced science advanced science & engineering & engineering  General Purpose Programming of Many-Core Devices and Systems : Why is parallel programming so hard? By Steven Ericsson-Zenith This presentation discusses problems and solutions to the general purpose programming problem on highly parallel systems. Prepared for the Gelato Many-Core Workshop in Shanghai, March 28th 2008. { Immediacy } AbstractGeneral Purpose Programming of Many-Core Devices
§1
This talk will highlight general purpose programming of many-core devices and many-core systems.
§2
General purpose programming of parallel machines is hard because the engineering models and languages we use
distract the programmer from the task at hand: the functional composition of applications. I will discuss
the process oriented programming model that is the basis of my work and also provide some historical
anecdotes from my experience with Occam and the Transputer.
§3
Parallel programming requires the use of formal methods to manage the complexities involved. The rules of
programming with concurrency are well-known, they have been known for decades, and they resolve the
traditional challenges. They simply must be followed. We can remove this burden from the programmer by
providing tools and systems built on rigorous foundations.
§4
I will also discuss the Carnap programming language and framework and the open source project that is
implementing a compiler for that language for many-core platforms.
§5
Further details of these technologies can be found on the web at Process-Interaction-Models.info and Carnap.info
| | ♦ |
Copyright © 2008, Steven Ericsson-Zenith (All Rights Reserved) Institute for Advanced Science & Engineering A 501(c)3 nonprofit organization incorporated in California. format: memeio/xsl/xhtml. this page generated by memeio Thursday, March 27th 2008 3:50 p.m. -08:00 |
|