Publication:
Phase asynchronous amr execution for productive and performant astrophysical flows

dc.contributor.coauthorTan Nguyen
dc.contributor.coauthorZhang, Weiqun
dc.contributor.coauthorAlmgren, Ann S.
dc.contributor.coauthorShalf, John
dc.contributor.departmentGraduate School of Sciences and Engineering
dc.contributor.kuauthorErten, Didem Unat
dc.contributor.kuauthorFarooqi, Muhammad Nufail
dc.contributor.schoolcollegeinstituteCollege of Engineering
dc.contributor.schoolcollegeinstituteGRADUATE SCHOOL OF SCIENCES AND ENGINEERING
dc.date.accessioned2024-11-09T23:05:22Z
dc.date.issued2018
dc.description.abstractAdaptive Mesh Refinement (AMR) is an approach to solving PDEs that reduces the computational and memory requirements at the expense of increased communication. Although adopting asynchronous execution can overcome communication issues, manually restructuring an AMR application to realize asynchrony is extremely complicated and hinders readability and long-term maintainability. To balance performance against productivity, we design a user-friendly API and adopt phase asynchronous execution model where all subgrids at an AMR level can be computed asynchronously. We apply the phase asynchrony to transform a real-world AMR application, CASTRO, which solves multicomponent compressible hydrodynamic equations for astrophysical flows. We evaluate the performance and programming effort required to use our carefully designed API and execution model for transitioning large legacy codes from synchronous to asynchronous execution up to 278,528 Intel-KNL cores. CASTRO is about 100K lines of code but less than 0.2% code changes are required to achieve significant performance improvement.
dc.description.indexedbyWOS
dc.description.openaccessNO
dc.description.sponsoredbyTubitakEuN/A
dc.description.sponsorshipTurkish Science and Technology Research Centre [215E185]
dc.description.sponsorshipOffice of Advanced Scientific Computing Research in the Department of Energy Office of Science [DE-AC02-05CH11231] Authors from Koc University are supported by the Turkish Science and Technology Research Centre Grant No: 215E185. Authors from Lawrence Berkeley National Laboratory were supported by the Office of Advanced Scientific Computing Research in the Department of Energy Office of Science under contract number DE-AC02-05CH11231.
dc.identifier.isbn978-1-5386-8384-2
dc.identifier.scopus2-s2.0-85064121632
dc.identifier.urihttps://hdl.handle.net/20.500.14288/8791
dc.identifier.wos494258800070
dc.keywordsAsynchronous runtime
dc.keywordsCommunication overlap
dc.language.isoeng
dc.publisherAssoc Computing Machinery
dc.relation.ispartofProceedings Of The International Conference For High Performance Computing, Networking, Storage, And Analysis (Sc'18)
dc.subjectComputer science
dc.subjectHardware architecture
dc.subjectEngineering
dc.subjectElectrical electronic engineering
dc.titlePhase asynchronous amr execution for productive and performant astrophysical flows
dc.typeConference Proceeding
dspace.entity.typePublication
local.contributor.kuauthorFarooqi, Muhammad Nufail
local.contributor.kuauthorErten, Didem Unat
local.publication.orgunit1GRADUATE SCHOOL OF SCIENCES AND ENGINEERING
local.publication.orgunit1College of Engineering
local.publication.orgunit2Graduate School of Sciences and Engineering
relation.isOrgUnitOfPublication3fc31c89-e803-4eb1-af6b-6258bc42c3d8
relation.isOrgUnitOfPublication.latestForDiscovery3fc31c89-e803-4eb1-af6b-6258bc42c3d8
relation.isParentOrgUnitOfPublication8e756b23-2d4a-4ce8-b1b3-62c794a8c164
relation.isParentOrgUnitOfPublication434c9663-2b11-4e66-9399-c863e2ebae43
relation.isParentOrgUnitOfPublication.latestForDiscovery8e756b23-2d4a-4ce8-b1b3-62c794a8c164

Files