agentleFS
Sign inStart free
SaddssSaddss/cursor-skills/skills/pp-separation-benchmark/SKILL.md

pp-separation-benchmark

Skill2 starsChanged 3 months agoLicence not identifiedView on GitHub

Benchmark an ingress-layer hot/cold pool separation scheme for LLM serving against a unified cache-aware baseline, and (optionally) reproduce it on real Kubernetes via kind. Use when the user asks to "测冷热分离"/"ingress split benchmark"/"分离 vs 统一哪个 QPS 高"/"扫冷热配比"/"pp separation"/"验证 ingress-nginx 冷热分流"/"在 kind 上跑真 vLLM 冷热池". Given a conversation-replay dataset with a known truncation rate, it (1) profiles the workload into hot (non-truncated) vs cold (truncated) request pools by request count AND token load, (2) measures the unified cache-aware baseline (B0) max sustainable QPS under a model SLO via binary search, (3) sweeps hot/cold GPU ratios (2h5c/3h4c/4h3c plus adaptive 1h6c/5h2c/6h1c) where the hot pool runs a cache-aware router + KV offload and the cold pool round-robins truncated requests, requiring each ratio to pass overall+hot+cold SLO and <=15% intra-pool imbalance, and (4) reports best ratio, QPS vs baseline, per-pool latency, internal/external cache hit rate, and imbalance. Includes a two-fidelity design: a "policy fidelity" tier (docker+nginx, zero network noise, answers "is separation worth it") and an "orchestration fidelity" tier (kind + ingress-nginx canary-by-header + NVIDIA device-plugin, answers "does this deploy on real K8s"). Carries hard-won fixes for docker-kill hangs, legitimately-failing ratios aborting the run, kind GPU passthrough via CDI without touching the host docker daemon, and glibc-vs-musl image gotchas.

  • Reads credentials

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 0 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

The licence could not be identified — read it at the source. We count its rules but don't quote them. Read them on GitHub.

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free