Imperial College London > Talks@ee.imperial > CAS Talks > POETS : partially ordered event triggered systems

POETS : partially ordered event triggered systems

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact George A Constantinides.

The POETS project is a five-year effort to build a combined software and hardware system which allows applications to be split into 1M+ concurrent state machines, and then to execute them on 100K+ concurrent hardware threads across 100+ tightly-coupled compute nodes. To achieve this we use an event-driven compute system with no global barriers or shared state, and re-write applications to use globally asynchronous algorithms. This talk will give an overview of the current hardware that has been built, and show how applications such as finite-volume solvers can be re-cast as an asynchronous system in order to run on such hardware. It will finish with initial strong and weak scaling results for a number of applications running on prototype systems with up to 1024 hardware threads.

This talk is part of the CAS Talks series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

Changes to Talks@imperial | Privacy and Publicity