000281767 001__ 281767
000281767 005__ 20250314084114.0
000281767 0247_ $$2doi$$a10.1145/2751205.2751216
000281767 0247_ $$2WOS$$aWOS:000493996800018
000281767 037__ $$aFZJ-2016-01449
000281767 041__ $$aEnglish
000281767 1001_ $$0P:(DE-HGF)0$$aShudler, Sergei$$b0
000281767 1112_ $$a29th International Conference on Supercomputing$$cNewport Beach$$d06/08/2015 - 06/11/2015$$gICS'15$$wCalifornia
000281767 245__ $$aExascaling Your Library
000281767 260__ $$aNew York, New York, USA$$bACM Press$$c2015
000281767 29510 $$aProceedings of the 29th ACM on International Conference on Supercomputing - ICS '15 - 2015. - ISBN 9781450335591
000281767 300__ $$a165-175
000281767 3367_ $$0PUB:(DE-HGF)8$$2PUB:(DE-HGF)$$aContribution to a conference proceedings$$bcontrib$$mcontrib$$s1454572096_31896
000281767 3367_ $$033$$2EndNote$$aConference Paper
000281767 3367_ $$2ORCID$$aCONFERENCE_PAPER
000281767 3367_ $$2DataCite$$aOutput Types/Conference Paper
000281767 3367_ $$2DRIVER$$aconferenceObject
000281767 3367_ $$2BibTeX$$aINPROCEEDINGS
000281767 520__ $$aMany libraries in the HPC field encapsulate sophisticated algorithms with clear theoretical scalability expectations. However, hardware constraints or programming bugs may sometimes render these expectations inaccurate or even plainly wrong. While algorithm engineers have already been advocating the systematic combination of analytical performance models with practical measurements for a very long time, we go one step further and show how this comparison can become part of automated testing procedures. The most important applications of our method include initial validation, regression testing, and benchmarking to compare implementation and platform alternatives. Advancing the concept of performance assertions, we verify asymptotic scaling trends rather than precise analytical expressions, relieving the developer from the burden of having to specify and maintain very fine grained and potentially non-portable expectations. In this way, scalability validation can be continuously applied throughout the whole development cycle with very little effort. Using MPI as an example, we show how our method can help uncover non-obvious limitations of both libraries and underlying platforms.
000281767 536__ $$0G:(DE-HGF)POF3-511$$a511 - Computational Science and Mathematical Methods (POF3-511)$$cPOF3-511$$fPOF III$$x0
000281767 536__ $$0G:(DE-Juel-1)ATMLPP$$aATMLPP - ATML Parallel Performance (ATMLPP)$$cATMLPP$$x1
000281767 588__ $$aDataset connected to CrossRef Conference
000281767 7001_ $$0P:(DE-HGF)0$$aCalotoiu, Alexandru$$b1
000281767 7001_ $$0P:(DE-HGF)0$$aHoefler, Torsten$$b2
000281767 7001_ $$0P:(DE-Juel1)140202$$aStrube, Alexandre$$b3$$ufzj
000281767 7001_ $$0P:(DE-HGF)0$$aWolf, Felix$$b4
000281767 773__ $$a10.1145/2751205.2751216
000281767 909CO $$ooai:juser.fz-juelich.de:281767$$pVDB
000281767 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)140202$$aForschungszentrum Jülich GmbH$$b3$$kFZJ
000281767 9131_ $$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$$x0
000281767 9141_ $$y2015
000281767 915__ $$0StatID:(DE-HGF)0550$$2StatID$$aNo Authors Fulltext
000281767 9201_ $$0I:(DE-Juel1)JSC-20090406$$kJSC$$lJülich Supercomputing Center$$x0
000281767 980__ $$acontrib
000281767 980__ $$aVDB
000281767 980__ $$aUNRESTRICTED
000281767 980__ $$aI:(DE-Juel1)JSC-20090406