MetaMorph: A Library Framework for Interoperable Kernels on Multi- and Many-Core Clusters

Simple item page
dc.contributor.authorHelal, A.en
dc.contributor.authorSathre, Paulen
dc.contributor.authorFeng, Wu-chunen
dc.contributor.departmentElectrical and Computer Engineeringen
dc.contributor.departmentComputer Scienceen
dc.date.accessioned2017-04-03T04:54:36Zen
dc.date.available2017-04-03T04:54:36Zen
dc.date.issued2016-11-15en
dc.description.abstractTo attain scalable performance efficiently, the HPC community expects future exascale systems to consist of multiple nodes, each with different types of hardware accelerators. In addition to GPUs and Intel MICs, additional candidate accelerators include embedded multiprocessors and FPGAs. End users need appropriate tools to efficiently use the available compute resources in such systems, both within a compute node and across compute nodes. As such, we present MetaMorph, a library framework designed to (automatically) extract as much computational capability as possible from HPC systems. Its design centers around three core principles: abstraction, interoperability, and adaptivity. To demonstrate its efficacy, we present a case study that uses the structured grids design pattern, which is heavily used in computational fluid dynamics. We show how MetaMorph significantly reduces the development time, while delivering performance and interoperability across an array of heterogeneous devices, including multicore CPUs, Intel MICs, AMD GPUs, and NVIDIA GPUs.en
dc.description.notesYes, full paper (Peer reviewed?)en
dc.description.versionPublished versionen
dc.format.extent119 - 129 (11) page(s)en
dc.format.mimetypeapplication/pdfen
dc.identifier.doihttps://doi.org/10.1109/SC.2016.10en
dc.identifier.isbn978-1-4673-8815-3en
dc.identifier.issn2167-4337en
dc.identifier.urihttp://hdl.handle.net/10919/76745en
dc.language.isoenen
dc.relation.ispartofACM/IEEE SC16: The International Conference for High Performance Computing, Networking, Storage and Analysis (also known as Supercomputing)en
dc.relation.urihttp://www.cs.vt.edu/~fengen
dc.rightsIn Copyrighten
dc.rights.urihttp://rightsstatements.org/vocab/InC/1.0/en
dc.subjectLibrariesen
dc.subjectHardwareen
dc.subjectInteroperabilityen
dc.subjectKernelen
dc.subjectPerformance Evaluationen
dc.subjectExascaleen
dc.subjectParallel Librariesen
dc.subjectPerformance Portabilityen
dc.subjectProgrammabilityen
dc.subjectAcceleratorsen
dc.subjectGPUen
dc.subjectMICen
dc.subjectCUDAen
dc.subjectOpenCLen
dc.subjectOpenMPen
dc.subjectMPIen
dc.subjectStructured Gridsen
dc.titleMetaMorph: A Library Framework for Interoperable Kernels on Multi- and Many-Core Clustersen
dc.title.serialACM/IEEE SC16: The International Conference for High Performance Computing, Networking, Storage and Analysis (also known as Supercomputing)en
dc.typeConference proceedingen
dc.type.dcmitypeTexten
pubs.finish-date2016-11-18en
pubs.organisational-group/Virginia Techen
pubs.organisational-group/Virginia Tech/All T&R Facultyen
pubs.organisational-group/Virginia Tech/Engineeringen
pubs.organisational-group/Virginia Tech/Engineering/COE T&R Facultyen
pubs.organisational-group/Virginia Tech/Engineering/Computer Scienceen
pubs.organisational-group/Virginia Tech/Faculty of Health Sciencesen
pubs.start-date2016-11-13en

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
helal-metamorph-sc16.pdf
Size:
671.65 KB
Format:
Adobe Portable Document Format
Description:
Submitted Version
Download
License bundle
Now showing 1 - 1 of 1
Name:
VTUL_Distribution_License_2016_05_09.pdf
Size:
18.09 KB
Format:
Adobe Portable Document Format
Description:
Download

Collections

All Faculty Deposits
Scholarly Works, Computer Science
Scholarly Works, Electrical and Computer Engineering