You got a big multicore, or manycore machine, but do not have a clue of how to actually use it, because your application doesn’t seem to scale naturally? Well, that seems to be a problem many people are facing in our new manycore age. One possible solution might be to accept less precise answers by relaxing synchronization constraints. That could allow us to circumvent Amdahl’s law when Gustafson is out of reach.
If you got ideas in that direction, let us know! We are organizing a workshop at SPLASH’12 called RACES.
Here a quote from the mission statement:
The RACES workshop is a first attempt at bringing a new school of thought to the attention of the computer science community: that of abandoning absolute certainty in favor of scalability and performance when considering parallel computation.
Today, multi core systems are becoming more and more the rule and conventional wisdom has been to scale up software for these systems by reducing synchronization constraints. Amdahl’s law however, implies that even the smallest fraction of inherently sequential code limits scaling. The RACES workshop wants to promote an approach towards scalability for many-core systems by reducing synchronization requirements drastically, possibly to the point of discarding them altogether. This will of course cause us to move from the certain into the merely probable.
Read more on: http://soft.vub.ac.be/races/