Cargando…

Parallel Compilation of CMS Software

LHC experiments have large amounts of software to build. CMS has studied ways to shorten project build times using parallel and distributed builds as well as improved ways to decide what to rebuild. We have experimented with making idle desktop and server machines easily available as a virtual build...

Descripción completa

Detalles Bibliográficos
Autores principales: Ashby, Shaun, Eulisse, Giulio, Schmid, Stefan, Tuura, Lassi A
Lenguaje:eng
Publicado: CERN 2004
Materias:
Acceso en línea:https://dx.doi.org/10.5170/CERN-2005-002.590
http://cds.cern.ch/record/803473
_version_ 1780904809213198336
author Ashby, Shaun
Eulisse, Giulio
Schmid, Stefan
Tuura, Lassi A
author_facet Ashby, Shaun
Eulisse, Giulio
Schmid, Stefan
Tuura, Lassi A
author_sort Ashby, Shaun
collection CERN
description LHC experiments have large amounts of software to build. CMS has studied ways to shorten project build times using parallel and distributed builds as well as improved ways to decide what to rebuild. We have experimented with making idle desktop and server machines easily available as a virtual build cluster using distcc and zeroconf. We have also tested variations of ccache and more traditional make dependency analysis. We report on our test results, with analysis of the factors that most improve or limit build performance.
id cern-803473
institution Organización Europea para la Investigación Nuclear
language eng
publishDate 2004
publisher CERN
record_format invenio
spelling cern-8034732019-09-30T06:29:59Zdoi:10.5170/CERN-2005-002.590http://cds.cern.ch/record/803473engAshby, ShaunEulisse, GiulioSchmid, StefanTuura, Lassi AParallel Compilation of CMS SoftwareDetectors and Experimental TechniquesLHC experiments have large amounts of software to build. CMS has studied ways to shorten project build times using parallel and distributed builds as well as improved ways to decide what to rebuild. We have experimented with making idle desktop and server machines easily available as a virtual build cluster using distcc and zeroconf. We have also tested variations of ccache and more traditional make dependency analysis. We report on our test results, with analysis of the factors that most improve or limit build performance.CERNCMS-CR-2004-051oai:cds.cern.ch:8034732004-11-12
spellingShingle Detectors and Experimental Techniques
Ashby, Shaun
Eulisse, Giulio
Schmid, Stefan
Tuura, Lassi A
Parallel Compilation of CMS Software
title Parallel Compilation of CMS Software
title_full Parallel Compilation of CMS Software
title_fullStr Parallel Compilation of CMS Software
title_full_unstemmed Parallel Compilation of CMS Software
title_short Parallel Compilation of CMS Software
title_sort parallel compilation of cms software
topic Detectors and Experimental Techniques
url https://dx.doi.org/10.5170/CERN-2005-002.590
http://cds.cern.ch/record/803473
work_keys_str_mv AT ashbyshaun parallelcompilationofcmssoftware
AT eulissegiulio parallelcompilationofcmssoftware
AT schmidstefan parallelcompilationofcmssoftware
AT tuuralassia parallelcompilationofcmssoftware