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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Massively parallel accelerators such as GPGPUs, manycores and FPGAs representa powerful and affordable tool for scientists who look to speed up simulationsof complex systems. However, porting code to such devices requires a detailedunderstanding of heterogeneous programming tools and effective strategies forparallelization. In this paper we present a source to source compilationapproach with whole-program analysis to automatically transform single-threadedFORTRAN 77 legacy code into OpenCL-accelerated programs with parallelizedkernels.  The main contributions of our work are: (1) whole-source refactoring to allowany subroutine in the code to be offloaded to an accelerator. (2) Minimizationof the data transfer between the host and the accelerator by eliminatingredundant transfers. (3) Pragmatic auto-parallelization of the code to beoffloaded to the accelerator by identification of parallelizable maps andreductions.  We have validated the code transformation performance of the compiler on theNIST FORTRAN 78 test suite and several real-world codes: the Large EddySimulator for Urban Flows, a high-resolution turbulent flow model; the shallowwater component of the ocean model Gmodel; the Linear Baroclinic Model, anatmospheric climate model and Flexpart-WRF, a particle dispersion simulator.  The automatic parallelization component has been tested on as 2-D ShallowWater model (2DSW) and on the Large Eddy Simulator for Urban Flows (UFLES) andproduces a complete OpenCL-enabled code base. The fully OpenCL-acceleratedversions of the 2DSW and the UFLES are resp. 9x and 20x faster on GPU than theoriginal code on CPU, in both cases this is the same performance as manuallyported code.

Domain-Specific Acceleration and Auto-Parallelization of Legacy Scientific Code in FORTRAN 77 using Source-to-Source Compilation