RTX 3090 mini-bench: ninfer cuts TTFT from 3.4s to 29ms, prompt processing ~76x faster
milkipedia · reddit · 2026-09-11
A rigorous RTX 3090 mini-benchmark compares the author's ninfer-3090 stack against llama.cpp on two Qwen models, with careful methodology (3 repeats, token-weighted medians, nonce-based cache busting, untimed warmups, server-side timing).
Results:
- Qwen3.6-35B-A3B: TTFT 29ms vs 3410ms; prompt processing 185,339 vs 2,440 tok/s (76x); generation 170.7 vs 148.6 tok/s.
- Qwen3.8-27B: TTFT 27ms vs 8160ms; PP 187,912 vs 1,013 tok/s; generation 34.9 vs 38.5 tok/s (slightly slower).
The headline: ninfer collapses time-to-first-token from seconds to tens of milliseconds and boosts prompt throughput by 1–2 orders of magnitude, transforming long-prompt interactivity; generation speed is a wash. The author published the full 7-prompt test set (400–12,900 tokens), configs (Ubuntu 24.04, GGUF IQ4XS/Q4KM baselines), and invites reproduction and critique.
More from Infra
- Ubicloud publishes "Fixing Tool-Call Accuracy in vLLM" — petrusenko_max · 2026-10-26
- Meta approves nonvoting Class C shares as 2026 capex guidance hits $135B–$145B — thione · 2026-10-26
- Unsloth Ships MiniMax-M3.1-Flash Quants: 2-bit Retains 91.5% Accuracy, Runs in 150GB RAM — danielhanchen · 2026-10-26
- Engineer with thousands of fabricated devices pours cold water on photonic crystal modulator yields — jwt0625 · 2026-10-26
- Anthropic S-1 reveals 48 cents of compute cost per dollar of Q2 revenue — nmasc_ · 2026-10-26
- Persistent AI joins UR network to route private AI inference with encrypted, IP-hidden traffic — 0xSammy · 2026-10-26