Output details
11 - Computer Science and Informatics
Lancaster University
Towards a holistic approach to auto-parallelization: integrating profile-driven parallelism detection and machine-learning based mapping
<08> Combines novel dynamic analysis and machine learning based parallelism mapping to develop a completely portable approach to auto-parallelisation, an important problem that has been outstanding for over 30 years. Published in PLDI as one of 41 papers accepted from 194 (21.1%) and won HiPEAC award in recognition of publication in a world-class conference. The significance of this work is reflected in wide citation, and it also included in a recently published book and taught in over 10 universities worldwide. Similar ideas have been adopted by industrial companies such as Vectorfabrics since this work had been published.