Publikace na Matematicko-fyzikální fakulta |
2011
Abstrakt
Actual trend set by CPU manufacturers and recent developement in the field of graphical processing units (GPUs) offered us the computational power of multi-core and many-core architectures. Database applications can benefit greatly from parallelism; however, many algorithms need to be redesigned and many technical issues need to be solved.
In this paper, we have focused on standard relational join problem from the perspective of current highly parallel architectures. We present comparison of different approaches and propose algorithm adaptations which can better utilize multiple computational cores.