During a recent talk, distributed computing researcher Ioan Raicu discussed how the challenges of exascale computing will be felt outside of the traditional supercomputing community as IaaS giants like Amazon move into the million-node range.
A new language could improve the quality of parallel code and automate some of the trickiest elements of multicore programming.
Researchers mitigate multicore challenges to refine current geological simulation capabilities.
While it may seem counterintuitive, workstations might see a resurgence due to cloud computing. Solutions to help users offload parallel processing tasks to the cloud while reaping the benefits of graphics and power of workstations are increasing, as are bigger, better desktop systems.
Portability or efficiency? Neither is guaranteed when writing explicit parallel code.