PoCL-R: A Scalable Low Latency Distributed OpenCL Runtime

Jan Solanti, Michal Babej, Julius Ikkala, Vinod Kumar Malamal Vadakital, Pekka Jääskeläinen

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

31 Downloads (Pure)


Offloading the most demanding parts of applications to an edge GPU server cluster to save power or improve the result quality is a solution that becomes increasingly realistic with new networking technologies. In order to make such a computing scheme feasible, an application programming layer that can provide both low latency and scalable utilization of remote heterogeneous computing resources is needed. To this end, we propose a latency-optimized scalable distributed heterogeneous computing runtime implementing the standard OpenCL API.

In the proposed runtime, network-induced latency is reduced by means of peer-to-peer data transfers and event synchronization as well as a streamlined control protocol implementation. Further improvements can be obtained streaming of source data directly from the producer device to the compute cluster. Compute cluster scalability is improved by distributing the command and event processing responsibilities to remote compute servers. We also show how a simple optional dynamic content size buffer OpenCL extension can significantly speed up applications that utilize variable length data.

For evaluation we present a smartphone-based augmented reality rendering case study which, using the runtime, receives 19× improvement in frames per second and 17× improvement in energy per frame when offloading parts of the rendering workload to a nearby GPU server. The remote kernel execution latency overhead of the runtime is only 60 ms on top of the network roundtrip time. The scalability on multi-server multi-GPU clusters is shown with a distributed large matrix multiplication application.
Original languageEnglish
Title of host publicationEmbedded Computer Systems: Architectures, Modeling, and Simulation
Subtitle of host publication21st International Conference, SAMOS 2021, Virtual Event, July 4–8, 2021, Proceedings
EditorsAlex Orailoglu, Matthias Jung, Marc Reichenbach
Number of pages17
ISBN (Electronic)978-3-031-04580-6
ISBN (Print)978-3-031-04579-0
Publication statusPublished - 2022
Publication typeA4 Article in conference proceedings
EventInternational Conference on Embedded Computer Systems: Architecture, Modeling and Simulation - Virtual event
Duration: 4 Jul 20218 Jul 2021

Publication series

NameLecture Notes in Computer Science
Volume13227 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


ConferenceInternational Conference on Embedded Computer Systems: Architecture, Modeling and Simulation
Abbreviated titleSAMOS XXI
Internet address

Publication forum classification

  • Publication forum level 1


Dive into the research topics of 'PoCL-R: A Scalable Low Latency Distributed OpenCL Runtime'. Together they form a unique fingerprint.

Cite this