Show simple item record

dc.contributor.authorSapio, Amedeo
dc.contributor.authorCanini, Marco
dc.contributor.authorHo, Chen-Yu
dc.contributor.authorNelson, Jacob
dc.contributor.authorKalnis, Panos
dc.contributor.authorKim, Changhoon
dc.contributor.authorKrishnamurthy, Arvind
dc.contributor.authorMoshref, Masoud
dc.contributor.authorPorts, Dan R. K.
dc.contributor.authorRichtarik, Peter
dc.date.accessioned2020-09-30T10:28:57Z
dc.date.available2019-02-27T05:38:27Z
dc.date.available2020-09-30T10:28:57Z
dc.date.issued2021
dc.identifier.urihttp://hdl.handle.net/10754/631179
dc.description.abstractTraining machine learning models in parallel is an increasingly important workload. We accelerate distributed parallel training by designing a communication primitive that uses a programmable switch dataplane to execute a key step of the training process. Our approach, SwitchML, reduces the volume of exchanged data by aggregating the model updates from multiple workers in the network. We co-design the switch processing with the end-host protocols and ML frameworks to provide an efficient solution that speeds up training by up to 5.5⇥ for a number of real-world benchmark models.
dc.relation.urlhttps://www.usenix.org/conference/nsdi21/presentation/sapio
dc.titleScaling Distributed Machine Learning with In-Network Aggregation
dc.typeConference Paper
dc.typeTechnical Report
dc.contributor.departmentComputer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
dc.contributor.departmentExtreme Computing Research Center
dc.contributor.departmentComputer Science Program
dc.contributor.departmentComputer Science
dc.conference.dateApril 12–14, 2021
dc.conference.name18th USENIX Symposium on Networked Systems Design and Implementation
dc.identifier.wosutWOS:000662976700045
dc.identifier.wosutWOS:000662976700045
dc.eprint.versionPre-print
dc.contributor.institutionMicrosoft
dc.contributor.institutionBarefoot Networks
dc.contributor.institutionUniversity of Washington
dc.identifier.pages785-808
dc.identifier.arxivid1903.06701
kaust.personSapio, Amedeo
kaust.personCanini, Marco
kaust.personHo, Chen-Yu
kaust.personKalnis, Panos
kaust.personRichtarik, Peter
refterms.dateFOA2019-02-27T05:38:28Z
dc.date.posted2019-02-27


Files in this item

Thumbnail
Name:
main.pdf
Size:
1.513Mb
Format:
PDF
Description:
Technical report updated version - released 2020-09-30

This item appears in the following Collection(s)

Show simple item record

VersionItemEditorDateSummary

*Selected version