Prefill and decode want opposite things from your router.
This is a AI post classified by Jev as AI infra & evals (a tutorial), kept by the AI Radar because it carries real work, not commentary.
Prefill and decode want opposite things from your router. Prefill needs to land on a pod that already holds its prefix, so it only computes the delta. Decode doesn't care, it can pull KV from any prefiller. So @_llm_d_ schedules each phase with its own profile. Greg Pereira (Red Hat AI, llm-d) breaks down why one routing decision isn't enough:
Posted by Red Hat AI (12.5k followers) 3 days ago · 48 likes · 3k views · view the original post on X. Kept by the AI Radar as AI infra & evals.
More AI work like this
- Trains your own large language model from scratch using plain PyTorch. — @tom_doerr
- Here's a useful use-case for @typesafeai for a change. — @iam_zachi
- This is genius. — @daniel_mac8
- Which GPU should you use for embedding workloads? — @runpod
- Big update to the @Alibaba_Qwen Qwen3.8-Flash-Next single DGX Spark recipe! — @jvr0x
- 🔰 The Databricks Advanced Learning Festival runs September 16 through October 14, with… — @databricks
- Q: What is a trace? — @HamelHusain
- Inference scaling part 1. — @rasbt
Every post is read and classified by Jev (TypeSafe): what it is, which market it belongs to, and whether the link is a real tool. 27.2k posts from 4.8k X accounts over the last 14 days, 1.2k tools, 19 markets. Collected every 5 minutes, fully re-ranked every hour — last update 2026-09-19 18:08 UTC. Full method.