Refactoring for introducing and tuning parallelism for heterogeneous multicore machines in Erlang
View/ Open
Date
24/06/2019Grant ID
EP/P020631/1
779882
644235
Keywords
Metadata
Show full item recordAbstract
This paper presents semi‐automatic software refactorings to introduce and tune structured parallelism in sequential Erlang code, as well as to generate code for running computations on GPUs and possibly other accelerators. Our refactorings are based on the lapedo framework for programming heterogeneous multi‐core systems in Erlang. lapedo is based on the PaRTE refactoring tool and also contains (1) a set of hybrid skeletons that target both CPU and GPU processors, (2) novel refactorings for introducing and tuning parallelism, and (3) a tool to generate the GPU offloading and scheduling code in Erlang, which is used as a component of hybrid skeletons. We demonstrate, on four realistic use‐case applications, that we are able to refactor sequential code and produce heterogeneous parallel versions that can achieve significant and scalable speedups of up to 220 over the original sequential Erlang program on a 24‐core machine with a GPU.
Citation
Janjic , V , Brown , C , Barwell , A & Hammond , K 2019 , ' Refactoring for introducing and tuning parallelism for heterogeneous multicore machines in Erlang ' , Concurrency and Computation : Practice and Experience , vol. Early View , e5420 , pp. 1-25 . https://doi.org/10.1002/cpe.5420
Publication
Concurrency and Computation : Practice and Experience
Status
Peer reviewed
DOI
10.1002/cpe.5420ISSN
1532-0634Type
Journal article
Description
This research has been generously supported by the European Union Framework 7 Para-Phrase project (IST-288570), EU Horizon 2020 projects RePhrase (H2020-ICT-2014-1), agreement number 644235; Teamplay (H2020-ICT 2017-1) agreement number 779882, and EPSRC Discovery, EP/P020631/1. EU COST Action IC1202: Timing Analysis On Code-Level (TACLe), and by a travel grant from EU HiPEAC.Collections
Items in the St Andrews Research Repository are protected by copyright, with all rights reserved, unless otherwise indicated.