English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

High-performance computing (HPC) is a major driver accelerating scientificresearch and discovery, from quantum simulations to medical therapeutics. Whilethe increasing availability of HPC resources is in many cases pivotal tosuccessful science, even the largest collaborations lack the computationalexpertise required for maximal exploitation of current hardware capabilities.The need to maintain multiple platform-specific codebases further complicatesmatters, potentially adding constraints on machines that can be utilized.Fortunately, numerous programming models are under development that aim tofacilitate portable codes for heterogeneous computing. One in particular isSYCL, an open standard, C++-based single-source programming paradigm. AmongSYCL's features is interoperability, a mechanism through which applications andthird-party libraries coordinate sharing data and execute collaboratively. Inthis paper, we leverage the SYCL programming model to demonstratecross-platform performance portability across heterogeneous resources. Wedetail our NVIDIA and AMD random number generator extensions to the oneMKLopen-source interfaces library. Performance portability is measured relative toplatform-specific baseline applications executed on four major hardwareplatforms using two different compilers supporting SYCL. The utility of ourextensions are exemplified in a real-world setting via a high-energy physicssimulation application. We show the performance of implementations thatcapitalize on SYCL interoperability are at par with native implementations,attesting to the cross-platform performance portability of a SYCL-basedapproach to scientific codes.

Achieving near native runtime performance and cross-platform performance portability for random number generation through SYCL interoperability