Thesis is the code of my thesis.
This thesis studies the feasibility of using GPUs for parallel processing algorithms applied to the filtering of events in a publish/subscribe system. This work involved a comparison between the experimental results of a sequential version (CPU version) and a parallel version of a referenced matching algorithm. This analysis has provided elements for arguing if any gains of using GPUs to perform fast matching of events are enough to offset the increased complexity of the process.
To run this code you need:
- CUDA 2.1 or superior
- C++ Compiler
- Of Course, a compatible NVIDIA Graphic Card