COMPUTER SCIENCE 2
- BSc/MSc/Diploma Thesis -
Model Checking on GPGPUs
A Comparative Study of CUDA Implementations for Sparse Matrix Vector
Multiplication in combination with Matrix Reordering Techniques to speed up CTMC
What is it all about?
Although Moore's Law is still correct, CPUs do not become much faster anymore. Instead of having a doubling of
clock rates every two years, nowadays we have multiple cores on one processor die. Consequently, sequential programs
do not speed up just "for free" with newer CPUs. Instead, new multi-core algorithms must be developed.
Even more challenging than multi-core processors (say, 2-16 cores) are massively parallel many-core systems (e.g.,