KVTide is a Kubernetes-native LLM serving runtime exploring cache-aware scheduling and proactive peer-to-peer KV mobility.
🏆[ICML 2026 Spotlight] Official implementation of "DecodeShare: Tracing the Shared Subspace of LLM Decode-Time Decisions"
A curated map of AFD, PD disaggregation, KV-cache systems, MoE serving, and re-aggregation baselines for LLM serving.
DiSpec — a from-scratch LLM inference engine: paged attention, continuous batching, CUDA-graph decode, speculative decoding, and prefill/decode disaggregation
A datacenter-scale simulation framework for energy- and SLO-aware LLM inference serving — non-invasive extension of CloudSim Plus
Measuring the KV-transfer tax of disaggregated prefill/decode LLM serving (vLLM + NIXL, 4x A10G): on PCIe-only hardware, disagg loses to plain data-parallel replication — measured, committed, reproducible.
Add a description, image, and links to the prefill-decode topic page so that developers can more easily learn about it.
To associate your repository with the prefill-decode topic, visit your repo's landing page and select "manage topics."