000872623 001__ 872623
000872623 005__ 20210130004236.0
000872623 0247_ $$2Handle$$a2128/23823
000872623 037__ $$aFZJ-2020-00115
000872623 1001_ $$0P:(DE-Juel1)169856$$aMorgenstern, Laura$$b0$$eCorresponding author$$ufzj
000872623 1112_ $$aErste Konferenz für ForschungssoftwareentwicklerInnen in Deutschland$$cPotsdam$$d2019-06-04 - 2019-06-06$$gdeRSE19$$wGermany
000872623 245__ $$aNUMA-Awareness as a Plug-In
000872623 260__ $$c2019
000872623 3367_ $$033$$2EndNote$$aConference Paper
000872623 3367_ $$2BibTeX$$aINPROCEEDINGS
000872623 3367_ $$2DRIVER$$aconferenceObject
000872623 3367_ $$2ORCID$$aCONFERENCE_POSTER
000872623 3367_ $$2DataCite$$aOutput Types/Conference Poster
000872623 3367_ $$0PUB:(DE-HGF)24$$2PUB:(DE-HGF)$$aPoster$$bposter$$mposter$$s1578897674_31747$$xAfter Call
000872623 520__ $$aMolecular dynamics (MD) has become a vital research method in biochemistry and materials science. GROMACS aims do develop a flexible and unified tool-box in the field of MD simulations. In MD, the fast multipole method (FMM) is used to compute all pairwise long-range interactions between N particles in time O(N). To tackle exascale, MD applications – as well as several other HPCapplications – have to target strong scaling. To meet the according requirements such as synchronization- and latency-awareness, software needs to adopt to specific hardware properties such ascaching and non-uniform memory access (NUMA). This poster shows how we added NUMA-awareness to our C++ tasking framework for fine-grained parallelism with an FMM as use case. However, the poster has an emphasis on separation of concerns through software architecture since the representation of NUMA in software is not only relevant for the FMM but should be reusable by similar applications.
000872623 536__ $$0G:(DE-HGF)POF3-511$$a511 - Computational Science and Mathematical Methods (POF3-511)$$cPOF3-511$$fPOF III$$x0
000872623 536__ $$0G:(DE-Juel1)PHD-NO-GRANT-20170405$$aPhD no Grant - Doktorand ohne besondere Förderung (PHD-NO-GRANT-20170405)$$cPHD-NO-GRANT-20170405$$x1
000872623 7001_ $$0P:(DE-Juel1)161429$$aHaensel, David$$b1$$ufzj
000872623 7001_ $$0P:(DE-Juel1)157750$$aBeckmann, Andreas$$b2$$ufzj
000872623 7001_ $$0P:(DE-Juel1)132152$$aKabadshow, Ivo$$b3$$ufzj
000872623 8564_ $$uhttps://juser.fz-juelich.de/record/872623/files/deRSE.pdf$$yOpenAccess
000872623 8564_ $$uhttps://juser.fz-juelich.de/record/872623/files/deRSE.pdf?subformat=pdfa$$xpdfa$$yOpenAccess
000872623 909CO $$ooai:juser.fz-juelich.de:872623$$pdriver$$pVDB$$popen_access$$popenaire
000872623 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)169856$$aForschungszentrum Jülich$$b0$$kFZJ
000872623 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)161429$$aForschungszentrum Jülich$$b1$$kFZJ
000872623 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)157750$$aForschungszentrum Jülich$$b2$$kFZJ
000872623 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)132152$$aForschungszentrum Jülich$$b3$$kFZJ
000872623 9131_ $$0G:(DE-HGF)POF3-511$$1G:(DE-HGF)POF3-510$$2G:(DE-HGF)POF3-500$$3G:(DE-HGF)POF3$$4G:(DE-HGF)POF$$aDE-HGF$$bKey Technologies$$lSupercomputing & Big Data$$vComputational Science and Mathematical Methods$$x0
000872623 9141_ $$y2019
000872623 915__ $$0StatID:(DE-HGF)0510$$2StatID$$aOpenAccess
000872623 9201_ $$0I:(DE-Juel1)JSC-20090406$$kJSC$$lJülich Supercomputing Center$$x0
000872623 9801_ $$aFullTexts
000872623 980__ $$aposter
000872623 980__ $$aVDB
000872623 980__ $$aUNRESTRICTED
000872623 980__ $$aI:(DE-Juel1)JSC-20090406