HeterPS: Distributed Deep Learning With Reinforcement Learning Based Scheduling in Heterogeneous Environments
Deep neural networks (DNNs) exploit many layers and a large number of parameters to achieve excellent performance. The training process of DNN models generally handles large-scale input data with many sparse features, which incurs high Input/Output (IO) cost, while some layers are compute-intensive....
Saved in:
Published in | arXiv.org |
---|---|
Main Authors | , , , , , , , , |
Format | Paper |
Language | English |
Published |
Ithaca
Cornell University Library, arXiv.org
07.06.2023
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Abstract | Deep neural networks (DNNs) exploit many layers and a large number of parameters to achieve excellent performance. The training process of DNN models generally handles large-scale input data with many sparse features, which incurs high Input/Output (IO) cost, while some layers are compute-intensive. The training process generally exploits distributed computing resources to reduce training time. In addition, heterogeneous computing resources, e.g., CPUs, GPUs of multiple types, are available for the distributed training process. Thus, the scheduling of multiple layers to diverse computing resources is critical for the training process. To efficiently train a DNN model using the heterogeneous computing resources, we propose a distributed framework, i.e., Paddle-Heterogeneous Parameter Server (Paddle-HeterPS), composed of a distributed architecture and a Reinforcement Learning (RL)-based scheduling method. The advantages of Paddle-HeterPS are three-fold compared with existing frameworks. First, Paddle-HeterPS enables efficient training process of diverse workloads with heterogeneous computing resources. Second, Paddle-HeterPS exploits an RL-based method to efficiently schedule the workload of each layer to appropriate computing resources to minimize the cost while satisfying throughput constraints. Third, Paddle-HeterPS manages data storage and data communication among distributed computing resources. We carry out extensive experiments to show that Paddle-HeterPS significantly outperforms state-of-the-art approaches in terms of throughput (14.5 times higher) and monetary cost (312.3% smaller). The codes of the framework are publicly available at: https://github.com/PaddlePaddle/Paddle. |
---|---|
AbstractList | Deep neural networks (DNNs) exploit many layers and a large number of parameters to achieve excellent performance. The training process of DNN models generally handles large-scale input data with many sparse features, which incurs high Input/Output (IO) cost, while some layers are compute-intensive. The training process generally exploits distributed computing resources to reduce training time. In addition, heterogeneous computing resources, e.g., CPUs, GPUs of multiple types, are available for the distributed training process. Thus, the scheduling of multiple layers to diverse computing resources is critical for the training process. To efficiently train a DNN model using the heterogeneous computing resources, we propose a distributed framework, i.e., Paddle-Heterogeneous Parameter Server (Paddle-HeterPS), composed of a distributed architecture and a Reinforcement Learning (RL)-based scheduling method. The advantages of Paddle-HeterPS are three-fold compared with existing frameworks. First, Paddle-HeterPS enables efficient training process of diverse workloads with heterogeneous computing resources. Second, Paddle-HeterPS exploits an RL-based method to efficiently schedule the workload of each layer to appropriate computing resources to minimize the cost while satisfying throughput constraints. Third, Paddle-HeterPS manages data storage and data communication among distributed computing resources. We carry out extensive experiments to show that Paddle-HeterPS significantly outperforms state-of-the-art approaches in terms of throughput (14.5 times higher) and monetary cost (312.3% smaller). The codes of the framework are publicly available at: https://github.com/PaddlePaddle/Paddle. |
Author | Liu, Ji Yao, Xuefeng Zhang, Minxu Yu, Dianhai Dou, Dejing Wu, Xinxuan Wu, Zhihua Ma, Yanjun Feng, Danlei |
Author_xml | – sequence: 1 givenname: Ji surname: Liu fullname: Liu, Ji – sequence: 2 givenname: Zhihua surname: Wu fullname: Wu, Zhihua – sequence: 3 givenname: Dianhai surname: Yu fullname: Yu, Dianhai – sequence: 4 givenname: Yanjun surname: Ma fullname: Ma, Yanjun – sequence: 5 givenname: Danlei surname: Feng fullname: Feng, Danlei – sequence: 6 givenname: Minxu surname: Zhang fullname: Zhang, Minxu – sequence: 7 givenname: Xinxuan surname: Wu fullname: Wu, Xinxuan – sequence: 8 givenname: Xuefeng surname: Yao fullname: Yao, Xuefeng – sequence: 9 givenname: Dejing surname: Dou fullname: Dou, Dejing |
BookMark | eNqNjM0KgkAYRYcoyMp3GGgtjGPaz7I0WrSIDFqK6adO5Dc1Pz1_FkHbVhfuOfeOSB8lQo84PAh8bzHjfEhcra-MMR7NeRgGDml3YEAd0hWNhTZKXKyBksYAd7qHXKHAmp6FaegRBFZSFdACmh9b57rz06KB0t7ehUD6uZQ1IEiraYJPoSS-Z3pCBlV-0-B-c0ym2-S02Xl3JR8WtMmu0irsUMYj5vvhcuGz4D_rBW4pS5Y |
ContentType | Paper |
Copyright | 2023. This work is published under http://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License. |
Copyright_xml | – notice: 2023. This work is published under http://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License. |
DBID | 8FE 8FG ABJCF ABUWG AFKRA AZQEC BENPR BGLVJ CCPQU DWQXO HCIFZ L6V M7S PIMPY PQEST PQQKQ PQUKI PRINS PTHSS |
DatabaseName | ProQuest SciTech Collection ProQuest Technology Collection Materials Science & Engineering Collection ProQuest Central (Alumni) ProQuest Central ProQuest Central Essentials ProQuest Central Technology Collection ProQuest One Community College ProQuest Central Korea SciTech Premium Collection ProQuest Engineering Collection Engineering Database Publicly Available Content Database ProQuest One Academic Eastern Edition (DO NOT USE) ProQuest One Academic ProQuest One Academic UKI Edition ProQuest Central China Engineering Collection |
DatabaseTitle | Publicly Available Content Database Engineering Database Technology Collection ProQuest Central Essentials ProQuest One Academic Eastern Edition ProQuest Central (Alumni Edition) SciTech Premium Collection ProQuest One Community College ProQuest Technology Collection ProQuest SciTech Collection ProQuest Central China ProQuest Central ProQuest Engineering Collection ProQuest One Academic UKI Edition ProQuest Central Korea Materials Science & Engineering Collection ProQuest One Academic Engineering Collection |
DatabaseTitleList | Publicly Available Content Database |
Database_xml | – sequence: 1 dbid: 8FG name: ProQuest Technology Collection url: https://search.proquest.com/technologycollection1 sourceTypes: Aggregation Database |
DeliveryMethod | fulltext_linktorsrc |
Discipline | Physics |
EISSN | 2331-8422 |
Genre | Working Paper/Pre-Print |
GroupedDBID | 8FE 8FG ABJCF ABUWG AFKRA ALMA_UNASSIGNED_HOLDINGS AZQEC BENPR BGLVJ CCPQU DWQXO FRJ HCIFZ L6V M7S M~E PIMPY PQEST PQQKQ PQUKI PRINS PTHSS |
ID | FETCH-proquest_journals_26011598103 |
IEDL.DBID | BENPR |
IngestDate | Thu Oct 10 18:15:43 EDT 2024 |
IsOpenAccess | true |
IsPeerReviewed | false |
IsScholarly | false |
Language | English |
LinkModel | DirectLink |
MergedId | FETCHMERGED-proquest_journals_26011598103 |
OpenAccessLink | https://www.proquest.com/docview/2601159810?pq-origsite=%requestingapplication% |
PQID | 2601159810 |
PQPubID | 2050157 |
ParticipantIDs | proquest_journals_2601159810 |
PublicationCentury | 2000 |
PublicationDate | 20230607 |
PublicationDateYYYYMMDD | 2023-06-07 |
PublicationDate_xml | – month: 06 year: 2023 text: 20230607 day: 07 |
PublicationDecade | 2020 |
PublicationPlace | Ithaca |
PublicationPlace_xml | – name: Ithaca |
PublicationTitle | arXiv.org |
PublicationYear | 2023 |
Publisher | Cornell University Library, arXiv.org |
Publisher_xml | – name: Cornell University Library, arXiv.org |
SSID | ssj0002672553 |
Score | 3.4788268 |
SecondaryResourceType | preprint |
Snippet | Deep neural networks (DNNs) exploit many layers and a large number of parameters to achieve excellent performance. The training process of DNN models generally... |
SourceID | proquest |
SourceType | Aggregation Database |
SubjectTerms | Artificial neural networks Computer networks Data communication Data storage Deep learning Distributed processing Machine learning Mathematical models Parameters Scheduling Training Workload |
Title | HeterPS: Distributed Deep Learning With Reinforcement Learning Based Scheduling in Heterogeneous Environments |
URI | https://www.proquest.com/docview/2601159810 |
hasFullText | 1 |
inHoldings | 1 |
isFullTextHit | |
isPrint | |
link | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwfV3dS8MwED_ciuCbn_gxR0Bfi02WdqkvwlxrETbKpri3kSbZ3INtXeurf7tJ6Jwg7DEcHMlxucvd_XIHcMs5yXjAlOtJP3CpkNjlPKQupopnIRaS2oruaBwkr_R55s-ahFvVwCo3NtEaalkIkyO_M62vtOtl2HsoP10zNcpUV5sRGi1wCKamTOsMonE6-c2ykKCv38y9f4bWeo_4EJyUl2p9BHsqP4Z9C7oU1Ql8JAaLkk7v0dC0rzWTp5REQ6VK1LQ9XaK3Vf2OJsr2NxU2lbelDbQHkmiqxS4NnnyJVjmyLAutFUqH9Cj6843tFG7i6OUxcTd7nDd6VM23p-6dQTsvcnUOKFQ6yPFZRhZMUkH0DfMWPPODkGDBQ84uoLOL0-Vu8hUcmJHqFg7V70C7Xn-pa-1466wLLRY_dRsZ69XoO_oB6p2P9Q |
link.rule.ids | 786,790,12792,21416,33406,33777,43633,43838 |
linkProvider | ProQuest |
linkToHtml | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwfV1LT8MwDLZgE4IbT_EYEAmuEX2kWcoFCbZSYJsmNsRuVZpkYwfaspb_TxJ1DAlp50hWYjl2bH_5DHDNuZdyyhR2ZEAxEdLFnIcEu0TxNHSFJLaj2x_Q-I08T4JJXXAra1jl0idaRy1zYWrkN4b6Sode5jp3xRc2U6NMd7UeobEJTeJT39g5ix5_aywebesXs__PzdrYEe1Cc8gLtdiDDZXtw5aFXIryAD5jg0QZjm5Rx5DXmrlTSqKOUgWqSU9n6H1efaBXZdlNhS3krdbudfyRaKSVLg2afIbmGbIic20TSif0qPvnE9shXEXd8UOMl3tMaisqk9WZ_SNoZHmmjgGFSqc4AUu9KZNEePp-OVOeBjT0XMFDzk6gtU7S6frlS9iOx_1e0nsavJzBjhmuboFR7RY0qsW3OtchuEovrJ5_AJ6Pj2U |
openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=HeterPS%3A+Distributed+Deep+Learning+With+Reinforcement+Learning+Based+Scheduling+in+Heterogeneous+Environments&rft.jtitle=arXiv.org&rft.au=Liu%2C+Ji&rft.au=Wu%2C+Zhihua&rft.au=Yu%2C+Dianhai&rft.au=Ma%2C+Yanjun&rft.date=2023-06-07&rft.pub=Cornell+University+Library%2C+arXiv.org&rft.eissn=2331-8422 |