Inference Infrastructure & ServingUCSD, PKU (Zhong et al.)
DistServe
DistServe — goodput-optimal PD disaggregation
Why it matters
Splits prefill and decode into separate GPU pools so that latency metrics can be tuned independently
Source — Primary sources first; secondary reports are labelled “Press”
Paperarxiv.orghttps://arxiv.org/abs/2401.09670Added in Week 35, 2026 · See that week