Strategies for regular segmented reductions on GPU

Publikation: Bidrag til bog/antologi/rapportKonferencebidrag i proceedingsForskningfagfællebedømt

We present and evaluate an implementation technique for regular segmented reductions on GPUs. Existing techniques tend to be either consistent in performance but relatively inefficient in absolute terms, or optimised for specific workloads and thereby exhibiting bad performance for certain input. We propose three different strategies for segmented reduction of regular arrays, each optimised for a particular workload. We demonstrate an implementation in the Futhark compiler that is able to employ all three strategies and automatically select the appropriate one at runtime. While our evaluation is in the context of the Futhark compiler, the implementation technique is applicable to any library or language that has a need for segmented reductions. We evaluate the technique on four microbenchmarks, two of which we also compare to implementations in the CUB library for GPU programming, as well as on two application benchmarks from the Rodinia suite. On the latter, we obtain speedups ranging from 1.3× to 1.7× over a previous implementation based on scans.

OriginalsprogEngelsk
TitelProceedings of the 6th ACM SIGPLAN International Workshop on Functional High-Performance Computing
Antal sider11
ForlagAssociation for Computing Machinery
Publikationsdato2017
Sider42-52
ISBN (Elektronisk)978-1-4503-5181-2
DOI
StatusUdgivet - 2017
Begivenhed6th ACM SIGPLAN International Workshop on Functional High-Performance Computing - Oxford, Storbritannien
Varighed: 7 sep. 20177 sep. 2017
Konferencens nummer: 6

Workshop

Workshop6th ACM SIGPLAN International Workshop on Functional High-Performance Computing
Nummer6
LandStorbritannien
ByOxford
Periode07/09/201707/09/2017

ID: 188403745