Files

T

noonghunnaandClaude Opus 4.7 1b9374b8b5 Add docs/FAQ.md — common questions answered for tweet click-throughs

Topics: hardware (4090 / 5090 / NVLink / non-NVIDIA / Windows-WSL2),
engine choice (vLLM vs llama.cpp / why not Ollama-LMStudio / MTP not
EAGLE / why not GGUF on vLLM / why AutoRound), performance (TPS
expectations, ctx-load decode drop, prefill cliffs explained,
vllm#40914), setup (model paths, GPU index override, multi-variant
ports, Open WebUI), community (bench contributions, bug reports,
Genesis bumping).

Linked from top-level README. Designed to absorb repeat issue-tracker
questions; each answer is 2-4 sentences with links to deeper docs.

Co-Authored-By: Claude Opus 4.7 (1M context) <[email protected]>

2026-04-28 22:17:54 +00:00

9.7 KiB

Raw Blame History

FAQ

Common questions about club-3090. If your question isn't here, open a GitHub Discussion — most things end up in this doc eventually.

Hardware

Can I use a 4090 instead of a 3090?

Yes — 4090 (Ada, sm_89) is strictly better than 3090 (Ampere, sm_86) for everything we ship. Slightly different kernel paths but no patches needed. Caveats: vLLM Genesis patches are tested on Ampere; tools should still work but TPS scaling is untested. Open an issue with numbers if you bench it.

Can I use a 5090?

Should work for vLLM (Blackwell adds new kernels but back-compat). The Marlin pad-sub-tile-n fork we mount targets Ampere edge cases — on Blackwell you can probably drop the /opt/ai/vllm-src/ mount. Not validated yet. We'd love numbers from a 5090 rig — use the Numbers from your rig issue template.

Do I need NVLink?

No. Our dual-card configs use PCIe-only, no NVLink. Custom all-reduce is disabled in the composes. NVLink would help dual-card TPS but it's not required, and the user has explicitly declined NVLink bridges as a default — adding the dependency would exclude most consumer rigs.

Does this work on AMD / Intel / Apple Silicon?

vLLM: NVIDIA-only (CUDA). llama.cpp: yes — pick the right Docker image (ghcr.io/ggml-org/llama.cpp:server-rocm for AMD, :server for CPU-only, or build from source for Apple Silicon). Update the image: line in the compose. The flags (--ngl, -fa on, --cache-type-k q4_0) work identically across backends.

Does this work on Windows / WSL2?

WSL2: yes, both engines. Make sure GPU passthrough is set up (nvidia-smi works inside WSL). Native Windows: vLLM doesn't support it; llama.cpp does — but use a native llama.cpp build, not Docker.

Engine choice

Why ship both vLLM and llama.cpp?

Different trades. vLLM is faster (51-89 TPS depending on config) and has full feature support (vision · tools · MTP spec-decode · streaming · reasoning), but its long-context configurations OOM on big tool returns (Cliff 1) or single prompts above 50-60K (Cliff 2). llama.cpp is slower (~21 TPS) but passes every stress test cleanly at full 262K context with vision and tools. For real-world tool-using agents (Claude Code, Cline, Hermes) that routinely send 25K+ tool messages, llama.cpp is the only single-card path that doesn't crash. See the launch frame: vLLM dual = max throughput, llama.cpp single = max robustness.

Why not Ollama?

Ollama wraps llama.cpp with a different model registry and slightly easier UX. It's fine for chat. Two reasons we don't ship it:

Ollama doesn't expose all llama.cpp flags we need (--cache-type-k q4_0, --mmproj, --spec-type ngram-mod, custom --parallel).
Ollama's model registry doesn't have the exact Unsloth GGUF quants we ship (UD-Q3_K_XL). You can run Ollama against an Unsloth GGUF manually, but at that point you've reimplemented our llama.cpp compose with a different wrapper.

Why not LM Studio?

LM Studio is GUI-driven and great for hobbyist use. We ship CLI/Docker because:

Reproducibility — pinned image SHAs + Genesis commit make exact bench runs across machines possible
Headless deployment — homelab racks, dev backends
Tool-call extraction across both engines on this exact model is non-trivial; LM Studio's defaults haven't been validated

Use LM Studio if you prefer a GUI and don't need the engineering. Use this repo if you want a tested config that another club-3090 user can match exactly.

Why MTP and not EAGLE?

We tried EAGLE — it's blocked on Qwen3-Next (the family Qwen3.5/3.6 belong to) by DeltaNet hybrid attention's lack of KV rollback support in vLLM/SGLang. MTP works because it's a different protocol (multi-token prediction at draft-head level, not a separate draft model). See INTERNALS.md "Speculative decoding" for the full forensic chain. Re-test triggers: if vllm#39931 lands or DeltaNet rollback support arrives upstream, EAGLE becomes viable again.

Why not GGUF on vLLM for this model?

Multiple gates blocked. Qwen3.6-27B GGUF on vLLM hits a chain of "fixed but-not-quite" issues — multimodal config routing, ParallelLMHead skip, the Qwen35TensorProcessor._reverse_reorder_v_heads weight loader producing garbage output on the 27B layout (transformers PR #45283 only validated on 0.8B). Tracked in INTERNALS.md. Use llama.cpp for GGUF on this model.

Why AutoRound INT4 not GPTQ / AWQ?

AutoRound (Lorbus) gave us +9% TPS over AWQ on this model. GPTQ has a similar quality bar but the AWQ + DFlash path failed (pad-Marlin × aux-layer interaction). AutoRound + Genesis + MTP is the production-validated path. AWQ is documented as a fallback for users who can't use AutoRound.

Performance

Why is single-card TPS lower than I expected?

Look at the TPS chart — single-card vLLM is 51-55 TPS narrative / 67-70 code at 48K, which beats most consumer-3090 numbers we've seen reported. If you're seeing materially lower, the most common causes are:

Power cap < 230 W (this rig benches at 230 W; 280 W gives ~+5%, 350 W ~+10%)
Wrong compose for your prompt shape (use vllm/fast-chat.yml for 20K chat, not long-vision.yml)
Genesis tree drift — git pull origin main between bench runs can change AL by ±15%. We pin to commit bf667c7 for this reason.

My TPS dropped after switching to 192K context. Why?

It shouldn't, much — we measured 50.93 TPS narr at 192K vs 50.53 at 32K (within variance) on long-vision.yml. If it dropped a lot, you're probably actually decoding into a long ctx (not just having KV pool reserved). Loaded-context decode is 2-4× cold short-prompt decode on any LLM. The TPS chart number is short-prompt cold; loaded numbers are in BENCHMARKS.md.

What's a "prefill cliff"?

VRAM-related OOM during prompt processing on single-card vLLM. Two cliffs documented:

Cliff 1 — TurboQuant tool-prefill OOM at ~138 MiB allocate request, fires on ≥25K tool messages at high mem-util (>0.95). Documented at single-3090 #1.
Cliff 2 — DeltaNet GDN forward OOM at ~50-60K single-prompt regardless of mem-util, fires on long single prompts. Both are in fla.ops and don't have file-replacement patches. Mitigation: vllm/default (48K + 0.92) for tools, dual-card or llama.cpp for long ctx.

vllm#40914 keeps coming up — what is it?

Sandermage's K+1 verify routing PR for vLLM. When it lands, the spec-verify cost we're paying on Ampere SM 8.6 (~22 TPS narrative regression vs pre-bug substrate) closes. Our v714 default on dev205 + Genesis v7.54 will jump from 51 narr to ~70 narr, matching what ampersandru measures on the older dev21 + v7.13 cascade-prone substrate. We track it in INTERNALS.md "Upstream tracker".

Setup

`bash scripts/setup.sh qwen3.6-27b` is downloading 20+ GB. Where does it go?

<repo>/models-cache/ by default. Override with MODEL_DIR=/path/to/your/scratch bash scripts/setup.sh qwen3.6-27b. See .env.example for all env vars.

My GPU isn't card 0 — how do I change it?

CUDA_VISIBLE_DEVICES=2 bash scripts/launch.sh --variant vllm/default (substitute your card index). For dual-card, pass two: CUDA_VISIBLE_DEVICES=2,3. The compose files inherit env from your shell.

Can I run multiple variants at once on the same machine?

You'd need different ports per variant. Edit ports: in the second compose (default is 8020:8000 → change to 8021:8000). Watch VRAM — two configs simultaneously typically don't fit on 24 GB.

Will this work behind Open WebUI?

Yes. Add a connection in Open WebUI's Settings → Connections → OpenAI: base URL http://localhost:8020/v1, any non-empty API key, model qwen3.6-27b-autoround. See docs/EXAMPLES.md.

Community / contribution

Can I add my benchmark numbers from a different rig?

Please do — open an issue using the Numbers from your rig template. We collect cross-rig data points in BENCHMARKS for community signal.

Found a bug — what should I include?

The bug report template asks for the data we always need: docker logs --tail 100, verify-full.sh output, nvidia-smi, your compose variant, and the repo commit. Skipping these means the first reply will just ask for them, costing you a round-trip.

How do I bump Genesis to a newer commit?

GENESIS_PIN=<new-commit-sha> bash scripts/setup.sh qwen3.6-27b and re-run bash scripts/verify-full.sh to confirm tools still work. Don't bump in production without re-running the verify suite — Genesis releases sometimes change spec-verify routing in ways that affect tool-call extraction.

9.7 KiB Raw Blame History Unescape Escape

FAQ