Suitability of the CBE for a scientific program


Jacob Jensen: Suitability of the CBE for a scientific program

The development of processors have reached a point of change. People who work with computer hardware and software will most likely have to adapt to new architectures in the near future. The Cell Broadband Engine (CBE) architecture is evidence of this change taking place.

This master thesis gives an overview of the CBE architecture focusing on the features which make it so different from today's standard computer architecture. Focus is also on the programming of the CBE and examples are given for those who wish to learn it. An analysis of two real world scientific programs, FDS and an Nbody simulation program, focusing on possibilities for porting them to the CBE, is then presented.

Eksaminer: Brian Vinter, DIKU
Censor: Josva Kleist, AAU