Ofan: Optimal Load Balancing for AI Training
Ofan, a switch-based destination-aware LB scheme that can reach $O(1)$ queueing, is suggested, a switch-based destination-aware LB scheme that can reach $O(1)$ queueing.
We have 2 of 5 papers
We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.
Not the right person? Other researchers publish under this name.
Ofan, a switch-based destination-aware LB scheme that can reach $O(1)$ queueing, is suggested, a switch-based destination-aware LB scheme that can reach $O(1)$ queueing.
This work shows that efficient offline-optimized routing enables efficient MoE training and inference on direct-connect topologies without the need for MoE traffic matrix or dynamic topology reconfiguration.
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.