000825383 001__ 825383
000825383 005__ 20221109161714.0
000825383 037__ $$aFZJ-2016-07846
000825383 041__ $$aEnglish
000825383 1001_ $$0P:(DE-Juel1)144723$$aDi Napoli, Edoardo$$b0$$ufzj
000825383 1112_ $$aJoint Laboratory for Extreme Scale Computing$$cKobe$$d2016-11-30 - 2016-12-02$$gJLESC$$wJapan
000825383 245__ $$aTowards Automated Load Balancing via Spectrum Slicing for FEAST-like solvers
000825383 260__ $$c2016
000825383 3367_ $$033$$2EndNote$$aConference Paper
000825383 3367_ $$2DataCite$$aOther
000825383 3367_ $$2BibTeX$$aINPROCEEDINGS
000825383 3367_ $$2DRIVER$$aconferenceObject
000825383 3367_ $$2ORCID$$aLECTURE_SPEECH
000825383 3367_ $$0PUB:(DE-HGF)6$$2PUB:(DE-HGF)$$aConference Presentation$$bconf$$mconf$$s1482342962_11191$$xAfter Call
000825383 520__ $$aSubspace iteration algorithms accelerated by rational filtering, such as FEAST, have recently re-emerged as a research topic in solving for interior eigenvalue problems. FEAST-like solvers are Rayleigh-Ritz solvers with rational filter functions, and as a result require re-orthogonalization on long vectors only in rare cases. Application of the filter functions, the computationally most expensive part, offers three levels of parallelism: 1) multiple spectral slices, 2) multiple linear system solves per slice, and 3) multiple right-hand sides per system solves. While the second and third level of parallelism are currently exploited, the first level is often difficult to efficiently realize.An efficient algorithmic procedure to load-balance multiple independent spectral slices is not yet available. Currently, existing solvers must rely on the user's prior knowledge. An automatic procedure to split a user specific interval into multiple load-balanced slices would greatly improve the state of the art. We outline how, both the algorithmic selection of filter functions and the spectral slices, can be at the center of load-balancing issues. Additionally, we present the tools and heuristics developed in an effort to tackle the problems.
000825383 536__ $$0G:(DE-HGF)POF3-511$$a511 - Computational Science and Mathematical Methods (POF3-511)$$cPOF3-511$$fPOF III$$x0
000825383 536__ $$0G:(DE-Juel1)SDLQM$$aSimulation and Data Laboratory Quantum Materials (SDLQM) (SDLQM)$$cSDLQM$$fSimulation and Data Laboratory Quantum Materials (SDLQM)$$x2
000825383 7001_ $$0P:(DE-Juel1)167415$$aWinkelmann, Jan$$b1$$eCorresponding author$$ufzj
000825383 909CO $$ooai:juser.fz-juelich.de:825383$$pVDB
000825383 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)144723$$aForschungszentrum Jülich$$b0$$kFZJ
000825383 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)167415$$aForschungszentrum Jülich$$b1$$kFZJ
000825383 9131_ $$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$$x0
000825383 9141_ $$y2016
000825383 915__ $$0StatID:(DE-HGF)0550$$2StatID$$aNo Authors Fulltext
000825383 9201_ $$0I:(DE-Juel1)JSC-20090406$$kJSC$$lJülich Supercomputing Center$$x0
000825383 9201_ $$0I:(DE-82)080012_20140620$$kJARA-HPC$$lJARA - HPC$$x1
000825383 980__ $$aconf
000825383 980__ $$aVDB
000825383 980__ $$aUNRESTRICTED
000825383 980__ $$aI:(DE-Juel1)JSC-20090406
000825383 980__ $$aI:(DE-82)080012_20140620