THE AI CONFERENCE


Associated tags: Cloud, Stanford University, LLM, VMS, Rama, Software, Research, Multiple inert gas elimination technique, Machine learning, SAN, Packet, NCCL, Conference, Medical device, Clockwork, Amazon

Clockwork to Present Two Talks on Speeding Up AI/ML Workloads by Eliminating Packet Delays/Loss at The AI Conference in San Francisco

Retrieved on: 
Monday, September 25, 2023
Cloud, Stanford University, LLM, VMS, Rama, Software, Research, Multiple inert gas elimination technique, Machine learning, SAN, Packet, NCCL, Conference, Medical device, Clockwork, Amazon

SAN FRANCISCO, Sept. 25, 2023 /PRNewswire/ -- Clockwork, a developer of advanced cloud technology for time-sensitive enterprise applications, today announced it will be delivering two technical presentations at the upcoming AI Conference in San Francisco on September 26-27. Clockwork's team will demonstrate how its two-step solution using Clockwork's Cloud Deluxe™ network application helps speed up AI/ML workloads, and improves cloud infrastructure performance.

Key Points: 
  • Clockwork's team will demonstrate how its two-step solution using Clockwork's Cloud Deluxe™ network application helps speed up AI/ML workloads, and improves cloud infrastructure performance.
  • Clockwork Senior Software Engineer Vinay Sriram will deliver a 25-minute talk titled "Speeding up AI/ML Workloads by Eliminating Packet Delays/Loss from 'Best Effort' TCP+Ethernet Networks" at 3 p.m. PT on Sept 26.
  • Sriram will also give a 10-minute lunchtime lightning talk titled "Eliminate TCP Packet Delays/Loss and Speed up AI/ML Workloads" at 12:45 p.m. PT on Sept 27.
  • He will also present results from NCCL tests and discuss how the technology performs on LLM workloads in different cloud environments.