Wed 7 Nov 2018 13:30 - 13:52 at Studio 2 - Language Design 1 Chair(s): Eelco Visser

This paper advocates programming high-performance code using partial evaluation. We present a clean-slate programming system with a simple, annotation-based, online partial evaluator that operates on a CPS-style intermediate representation. Our system exposes code generation for accelerators (vectorization/parallelization for CPUs and GPUs) via compiler-known higher-order functions that can be subjected to partial evaluation. This way, generic implementations can be instantiated with target-specific code at compile time.

In our experimental evaluation we present three extensive case studies from image processing, ray tracing, and genome sequence alignment. We demonstrate that using partial evaluation, we obtain high-performance implementations for CPUs and GPUs from one language and one code base in a generic way. The performance of our codes is mostly within 10%, often closer to the performance of multi man-year, industry-grade, manually-optimized expert codes that are considered to be among the top contenders in their fields.

Wed 7 Nov
Times are displayed in time zone: Guadalajara, Mexico City, Monterrey change

13:30 - 15:00: Language Design 1OOPSLA at Studio 2
Chair(s): Eelco VisserDelft University of Technology
13:30 - 13:52
Talk
OOPSLA
Roland LeißaSaarland University, Germany, Klaas BoescheSaarland University, Sebastian HackSaarland University, Germany, Arsène Pérard-GayotSaarland University, Germany, Richard MembarthDFKI, Germany, Philipp SlusallekDFKI, Germany, André MüllerJohannes Gutenberg University, Bertil SchmidtJohannes Gutenberg University
13:52 - 14:15
Talk
OOPSLA
Jeff BezansonJulia Computing, Benjamin ChungNortheastern University, Jiahao ChenCapital One, Stefan Karpinski, Viral B ShahJulia Computing, Jan VitekNortheastern University, Lionel ZoubritzkyÉcole Normale Supérieure
14:15 - 14:37
Talk
OOPSLA
14:37 - 15:00
Talk
OOPSLA