High Performance Computing Group - Industrial Affiliates Program

Multiprocessing Applications

Projects



Parallel Database Algorithms

This project deals with the design and evaluation of parallel database algorithms. The aim of this project is to show how locality exploitation can improve the execution of database algorithms specially for in-core text databases. Our first approach has been related to sorting in-core algorithms which are a key point in many database areas and specially in text retrieval. We have proposed an algorithm based on Radix sort which reduces the communication in the number of steps and amount of data communicated.
...clik here for more information...
Systematic mapping of algorithms to hardware
One method to execute efficiently parallel algorithms with hypercube communication topology (i.e. FFT algorithms, All-to-All personalized communications, Jacobi methods for singular value decomposition and eigenvalue computation) on mesh and torus multicomputers has been developed. It was proposed and evaluated a new embedding of hypercubes on torus multicomputers. It was shown that this embedding is optimal for rings in terms of execution time. In addition, some new techniques to reduce the communication costs inhypercube algorithms have been studied.
...clik here for more information...