Acceleration of computations in AI REML for single-step GBLUP models.

MASUDA, Y. - AGUILAR, I. - TSURUTA, S. - MISZTAL, I.

Resumen:

ABSTRACT.The objective of this study was to evaluate the advantage of the YAMS package over the FSPAK package in average-information (AI) REML for single-step GBLUP models. Data sets from broiler and Holsteins were used in this study. (Co)variance components were estimated with the AIREMLF90 program which could switch YAMS and FSPAK for sparse operations. The YAMS package used the BLAS and LAPACK libraries using all the 16 cores on CPU. For a single-trait model applied to the data contained over 15,000 genotyped animals, FSPAK took over 4 hours to finish the first 5 rounds while YAMS took 20 minutes. For a 4-trait model applied to the same data set, FSPAK failed in the sparse factorization while YAMS took 5 hours to finish the first 5 rounds. The use of YAMS can dramatically increase speed and stability of AIREMLF90 for single-step GBLUP models.


Detalles Bibliográficos
2014
Single step GBLUP
Supernodal methods
Variance component estimation
Inglés
Instituto Nacional de Investigación Agropecuaria
AINFO
http://www.ainfo.inia.uy/consulta/busca?b=pc&id=61923&biblioteca=vazio&busca=61923&qFacets=61923
Acceso abierto
_version_ 1805580521366880256
author MASUDA, Y.
author2 AGUILAR, I.
TSURUTA, S.
MISZTAL, I.
author2_role author
author
author
author_facet MASUDA, Y.
AGUILAR, I.
TSURUTA, S.
MISZTAL, I.
author_role author
bitstream.checksum.fl_str_mv 9d8b64881ad13799cc1c20c9d2d714ba
bitstream.checksumAlgorithm.fl_str_mv MD5
bitstream.url.fl_str_mv https://redi.anii.org.uy/jspui/bitstream/20.500.12381/1732/1/sword-2022-10-20T22%3a49%3a01.original.xml
collection AINFO
dc.creator.none.fl_str_mv MASUDA, Y.
AGUILAR, I.
TSURUTA, S.
MISZTAL, I.
dc.date.accessioned.none.fl_str_mv 2022-10-21T01:49:01Z
dc.date.available.none.fl_str_mv 2022-10-21T01:49:01Z
dc.date.issued.none.fl_str_mv 2014
dc.date.updated.none.fl_str_mv 2022-10-21T01:49:01Z
dc.description.abstract.none.fl_txt_mv ABSTRACT.The objective of this study was to evaluate the advantage of the YAMS package over the FSPAK package in average-information (AI) REML for single-step GBLUP models. Data sets from broiler and Holsteins were used in this study. (Co)variance components were estimated with the AIREMLF90 program which could switch YAMS and FSPAK for sparse operations. The YAMS package used the BLAS and LAPACK libraries using all the 16 cores on CPU. For a single-trait model applied to the data contained over 15,000 genotyped animals, FSPAK took over 4 hours to finish the first 5 rounds while YAMS took 20 minutes. For a 4-trait model applied to the same data set, FSPAK failed in the sparse factorization while YAMS took 5 hours to finish the first 5 rounds. The use of YAMS can dramatically increase speed and stability of AIREMLF90 for single-step GBLUP models.
dc.identifier.none.fl_str_mv http://www.ainfo.inia.uy/consulta/busca?b=pc&id=61923&biblioteca=vazio&busca=61923&qFacets=61923
dc.language.iso.none.fl_str_mv en
eng
dc.rights.es.fl_str_mv Acceso abierto
dc.rights.none.fl_str_mv info:eu-repo/semantics/openAccess
dc.source.none.fl_str_mv reponame:AINFO
instname:Instituto Nacional de Investigación Agropecuaria
instacron:Instituto Nacional de Investigación Agropecuaria
dc.subject.none.fl_str_mv Single step GBLUP
Supernodal methods
Variance component estimation
dc.title.none.fl_str_mv Acceleration of computations in AI REML for single-step GBLUP models.
dc.type.none.fl_str_mv ConferenceObject
PublishedVersion
info:eu-repo/semantics/conferenceObject
dc.type.version.none.fl_str_mv info:eu-repo/semantics/publishedVersion
description ABSTRACT.The objective of this study was to evaluate the advantage of the YAMS package over the FSPAK package in average-information (AI) REML for single-step GBLUP models. Data sets from broiler and Holsteins were used in this study. (Co)variance components were estimated with the AIREMLF90 program which could switch YAMS and FSPAK for sparse operations. The YAMS package used the BLAS and LAPACK libraries using all the 16 cores on CPU. For a single-trait model applied to the data contained over 15,000 genotyped animals, FSPAK took over 4 hours to finish the first 5 rounds while YAMS took 20 minutes. For a 4-trait model applied to the same data set, FSPAK failed in the sparse factorization while YAMS took 5 hours to finish the first 5 rounds. The use of YAMS can dramatically increase speed and stability of AIREMLF90 for single-step GBLUP models.
eu_rights_str_mv openAccess
format conferenceObject
id INIAOAI_b05a9fda2a27c569e92f1824505fdff9
instacron_str Instituto Nacional de Investigación Agropecuaria
institution Instituto Nacional de Investigación Agropecuaria
instname_str Instituto Nacional de Investigación Agropecuaria
language eng
language_invalid_str_mv en
network_acronym_str INIAOAI
network_name_str AINFO
oai_identifier_str oai:redi.anii.org.uy:20.500.12381/1732
publishDate 2014
reponame_str AINFO
repository.mail.fl_str_mv lorrego@inia.org.uy
repository.name.fl_str_mv AINFO - Instituto Nacional de Investigación Agropecuaria
repository_id_str
rights_invalid_str_mv Acceso abierto
spelling 2022-10-21T01:49:01Z2022-10-21T01:49:01Z20142022-10-21T01:49:01Zhttp://www.ainfo.inia.uy/consulta/busca?b=pc&id=61923&biblioteca=vazio&busca=61923&qFacets=61923ABSTRACT.The objective of this study was to evaluate the advantage of the YAMS package over the FSPAK package in average-information (AI) REML for single-step GBLUP models. Data sets from broiler and Holsteins were used in this study. (Co)variance components were estimated with the AIREMLF90 program which could switch YAMS and FSPAK for sparse operations. The YAMS package used the BLAS and LAPACK libraries using all the 16 cores on CPU. For a single-trait model applied to the data contained over 15,000 genotyped animals, FSPAK took over 4 hours to finish the first 5 rounds while YAMS took 20 minutes. For a 4-trait model applied to the same data set, FSPAK failed in the sparse factorization while YAMS took 5 hours to finish the first 5 rounds. The use of YAMS can dramatically increase speed and stability of AIREMLF90 for single-step GBLUP models.https://hdl.handle.net/20.500.12381/1732enenginfo:eu-repo/semantics/openAccessAcceso abiertoSingle step GBLUPSupernodal methodsVariance component estimationAcceleration of computations in AI REML for single-step GBLUP models.ConferenceObjectPublishedVersioninfo:eu-repo/semantics/conferenceObjectinfo:eu-repo/semantics/publishedVersionreponame:AINFOinstname:Instituto Nacional de Investigación Agropecuariainstacron:Instituto Nacional de Investigación AgropecuariaMASUDA, Y.AGUILAR, I.TSURUTA, S.MISZTAL, I.SWORDsword-2022-10-20T22:49:01.original.xmlOriginal SWORD entry documentapplication/octet-stream1869https://redi.anii.org.uy/jspui/bitstream/20.500.12381/1732/1/sword-2022-10-20T22%3a49%3a01.original.xml9d8b64881ad13799cc1c20c9d2d714baMD5120.500.12381/17322022-10-20 22:49:01.475oai:redi.anii.org.uy:20.500.12381/1732Gobiernohttp://inia.uyhttps://redi.anii.org.uy/oai/requestlorrego@inia.org.uyUruguayopendoar:2022-10-21T01:49:01AINFO - Instituto Nacional de Investigación Agropecuariafalse
spellingShingle Acceleration of computations in AI REML for single-step GBLUP models.
MASUDA, Y.
Single step GBLUP
Supernodal methods
Variance component estimation
status_str publishedVersion
title Acceleration of computations in AI REML for single-step GBLUP models.
title_full Acceleration of computations in AI REML for single-step GBLUP models.
title_fullStr Acceleration of computations in AI REML for single-step GBLUP models.
title_full_unstemmed Acceleration of computations in AI REML for single-step GBLUP models.
title_short Acceleration of computations in AI REML for single-step GBLUP models.
title_sort Acceleration of computations in AI REML for single-step GBLUP models.
topic Single step GBLUP
Supernodal methods
Variance component estimation
url http://www.ainfo.inia.uy/consulta/busca?b=pc&id=61923&biblioteca=vazio&busca=61923&qFacets=61923

Resultados similares