Choose timezone
Your profile timezone:
Powered by Indico
v3.3.3-pre
Future computer systems will build on a many core foundation. The quest to optimize performance per watt will push these cores to specialize, so these will be heterogeneous many core systems. The hardware trends are clear; it's the software response that is a bit murky. Software must change to take advantage of these new platforms. This is a technical problem, but it's also a social problem as the software must address the needs of different application and programmer communities.
Standards such as OpenCL are a critical part of the solution. This new programming language lets programmers write software that runs on CPUs, GPUs or combinations thereof. A long term solution, however, will only emerge from a clear understanding of how programmers architect parallel software. This is the focus of my research on Design Patterns carried out in collaboration with the ParLab at UC Berkeley. In this talk, I will describe these two projects and how they work together to address many core programming challenges.
About the speaker
Tim Mattson is an applications programmer (Ph.D. Chemistry, UCSC, 1985). He does linear algebra, finds oil, shakes molecules, solves differential equations, and models electrons in simple atomic systems. He has spent his career working with computer scientists to make sure the needs of parallel applications programmers are met.
Dr. Mattson is also engaged in a long term research program with UC Berkeley's ParLab on abstractions that bridge across parallel system design, parallel programming environments, and application software. This work builds on his book "Design Patterns in Parallel Programming" (written with Professors Beverly Sanders and Berna Massingill). The patterns provide the "human angle" and help keep his research focused on technologies that help general programmers solve real problems.
Organised by: S.Jarp and Miguel Angel Marquina - IT Department
CERN Computing Seminars and Colloquia