Case Studies in Program Design for Heterogeneous Parallelism. Sekerinski, E. & Zhang, T. Technical Report CAS-14-03-ES, McMaster University, March, 2014.
Case Studies in Program Design for Heterogeneous Parallelism [pdf]Paper  abstract   bibtex   
This collection is the outcome of students projects of the graduate course CAS 766 Concurrent Programming at McMaster University in the fall of 2013. The projects are on heteroge- neous computing, the execution of program on a combination of CPUs, FPGAs and GPUs. With processor frequencies having reached a ceiling (for a number of years already!) and power consumption limiting the complexity of single (multi-core) CPUs, heterogeneous computing has the promise of speeding up computations through parallelism while reducing power consumption. The last point is essential for everything from mobile devices to server farms. However, heterogeneous computing requires a different programming model than the traditional models of shared variables and message passing.

Downloads: 0