Automatic Parallelization and Locality Optimization of Beamforming Algorithms
Résumé
This paper demonstrates the benefits of a global optimization strategy using a new automatic parallelization and locality optimization methodology for high performance embedded computing algorithms that occur in adaptive radar systems, for modern multi-core computing chips. As a baseline, the resulting performance was compared against the performance that could be obtained using highly optimized math libraries.
Origine : Fichiers produits par l'(les) auteur(s)
Loading...