kth.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Toward GPU-centric Networking on Commodity Hardware
KTH, School of Electrical Engineering and Computer Science (EECS), Computer Science, Software and Computer systems, SCS.ORCID iD: 0000-0002-9400-324X
KTH, School of Electrical Engineering and Computer Science (EECS), Computer Science, Software and Computer systems, SCS.ORCID iD: 0000-0002-9780-873X
KTH, School of Electrical Engineering and Computer Science (EECS), Computer Science, Communication Systems, CoS.ORCID iD: 0000-0002-6066-746X
KTH, School of Electrical Engineering and Computer Science (EECS), Computer Science, Software and Computer systems, SCS.ORCID iD: 0000-0002-1256-1070
2024 (English)In: 7th International Workshop on Edge Systems, Analytics and Networking (EdgeSys 2024),  April 22, 2024, Athens, Greece, New York: ACM Digital Library, 2024Conference paper, Published paper (Refereed)
Abstract [en]

GPUs are emerging as the most popular accelerator for many applications, powering the core of machine learning applications. In networked GPU-accelerated applications input & output data typically traverse the CPU and the OS network stack multiple times, getting copied across the system’s main memory. These transfers increase application latency and require expensive CPU cycles, reducing the system’s efficiency, and increasing the overall response times. These inefficiencies become of greater importance in latency-bounded deployments, or with high throughput, where copy times could quickly inflate the response time of modern GPUs.We leverage the efficiency and kernel-bypass benefits of RDMA to transfer data in and out of GPUs without using any CPU cycles or synchronization. We demonstrate the ability of modern GPUs to saturate a 100-Gbps link, and evaluate the network processing timein the context of an inference serving application.

Place, publisher, year, edition, pages
New York: ACM Digital Library, 2024.
Keywords [en]
GPUs, Commodity Hardware, Inference Serving, RDMA
National Category
Communication Systems Computer Systems
Research subject
Computer Science
Identifiers
URN: urn:nbn:se:kth:diva-345624DOI: 10.1145/3642968.3654820OAI: oai:DiVA.org:kth-345624DiVA, id: diva2:1851484
Conference
7th International Workshop on Edge Systems, Analytics and Networking (EdgeSys 2024), April 22, 2024, Athens, Greece 
Note

QC 20240415

Part of ISBN 979-8-4007-0539-7

Available from: 2024-04-15 Created: 2024-04-15 Last updated: 2024-04-15Bibliographically approved

Open Access in DiVA

rdma-from-gpu-edgesys24(2942 kB)125 downloads
File information
File name FULLTEXT01.pdfFile size 2942 kBChecksum SHA-512
11ac62beed24dbee3f2d1c82d2bde875bb6ec32e0bc4aaa0d763311f6a09f4bd3e00ed17bec58b74dadd36d41683343e4c90d4ebe59837521accf7348d1b698f
Type fulltextMimetype application/pdf

Other links

Publisher's full text

Authority records

Girondi, MassimoScazzariello, MarianoMaguire Jr., Gerald Q.Kostic, Dejan

Search in DiVA

By author/editor
Girondi, MassimoScazzariello, MarianoMaguire Jr., Gerald Q.Kostic, Dejan
By organisation
Software and Computer systems, SCSCommunication Systems, CoS
Communication SystemsComputer Systems

Search outside of DiVA

GoogleGoogle Scholar
Total: 126 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

doi
urn-nbn

Altmetric score

doi
urn-nbn
Total: 304 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf