Low pain vs no pain multicore Haskells

TY - GEN

T1 - Low pain vs no pain multicore Haskells

AU - Aswad , Mustafa

AU - Trinder , Phil

AU - Al Zain, Abdallah

AU - Michaelson , Greg

AU - Berthold, Jost

N1 - Conference code: 10

PY - 2010

Y1 - 2010

N2 - Multicores are becoming the dominant processor technology andfunctional languages are theoretically well suited to exploit them. In practice,however, implementing effective high level parallel functional languages is extremelychallenging.This paper is the first programming and performance comparison of functionalmulticore technologies and reports some of the first ever multicore resultsfor two languages. As such it reflects the growing maturity of the field bysystematically evaluating four parallel Haskell implementations on a commonmulticore architecture. The comparison contrasts the programming effort eachlanguage requires with the parallel performance delivered. The study uses 15’typical’ programs to compare a ’no pain’, i.e. entirely implicit, parallel languagewith three ’low pain’, i.e. semi-explicit languages.The parallel Haskell implementations use different versions of GHC compilertechnology, and hence the comparative performance metric is speedup whichnormalises against sequential performance. We ground the speedup comparisonsby reporting both sequential and parallel runtimes and efficiencies for three ofthe languages. Our experiments focus on the number of programs improved, theabsolute speedups delivered, the parallel scalability, and the program changesrequired to coordinate parallelism. The results are encouraging and, on occasion,surprising.

AB - Multicores are becoming the dominant processor technology andfunctional languages are theoretically well suited to exploit them. In practice,however, implementing effective high level parallel functional languages is extremelychallenging.This paper is the first programming and performance comparison of functionalmulticore technologies and reports some of the first ever multicore resultsfor two languages. As such it reflects the growing maturity of the field bysystematically evaluating four parallel Haskell implementations on a commonmulticore architecture. The comparison contrasts the programming effort eachlanguage requires with the parallel performance delivered. The study uses 15’typical’ programs to compare a ’no pain’, i.e. entirely implicit, parallel languagewith three ’low pain’, i.e. semi-explicit languages.The parallel Haskell implementations use different versions of GHC compilertechnology, and hence the comparative performance metric is speedup whichnormalises against sequential performance. We ground the speedup comparisonsby reporting both sequential and parallel runtimes and efficiencies for three ofthe languages. Our experiments focus on the number of programs improved, theabsolute speedups delivered, the parallel scalability, and the program changesrequired to coordinate parallelism. The results are encouraging and, on occasion,surprising.

M3 - Article in proceedings

SN - 978-1-84150-158-1

SP - 49

EP - 64

BT - Trends in Functional Programming Volume 10

A2 - Horváth, Zoltán

A2 - Zsók, Viktória

A2 - Achten, Peter

A2 - Koopman, Pieter

PB - Intellect

T2 - 10th Symposium on Trends in Functional Programming

Y2 - 2 June 2009 through 4 June 2009

ER -