single source · inference & serving
Zero-copy fused token permutation and unpermutation
It sends tokens directly to their expert-grouped positions on remote ranks, eliminating intermediate copies.1 so tokens are sent directly to their expert-grouped positions on remote ranks, and views of the communication buffer are returned directly to the computation, eliminating intermediate copies.2
How this definition was written
Summarised by glm-5.3-flash from 1 source; every sentence is checked against the quote it cites.
We implement a fused per- mute/unpermute operator in which the planning kernel precomputes the destination of every token
— k3 tech reportso tokens are sent directly to their expert-grouped positions on remote ranks, and views of the communication buffer are returned directly to the computation, eliminating intermediate copies
— k3 tech report
Also called zero-copy communication, fused per-mute/unpermute operator.
Filed under inference & serving › serving parallelism