Builds / 2x RTX 3090
2x RTX 3090: Qwen3.8 27B W4A16 AutoRound on vLLM TP2, 91.2 tok/s prose (Tech2wild)
Last updated:
Source: github.com ↗More source: github.com ↗📏 How measured: reported by the original source; RigBase re-checked the numbers at that source
- GPU(s)
- 2x RTX 3090
Amazon · eBay (used) · Newegg - VRAM / unified memory (GB)
- 48
- CPU
- not stated
- Motherboard
- not stated
- RAM
- not stated
- PSU
- not stated
- Other
- 2× RTX 3090 (Ampere)
- Reported total cost
- not stated
- Power
- not stated
Measured speed (generation tokens/sec)
| Model | Quant | Engine | tok/s | Prompt processing (tok/s) | Details |
|---|---|---|---|---|---|
| Qwen3.8 27B | W4A16 AutoRound + DFlash2 | vLLM TP2, 1 stream | 91.2 | not stated | ⚡ spec-decode (DFlash)📏 Single-stream prose decode as published in Tech2wild's recipe README (free-prose decode, single stream). |
Numbers by Tech2wild (tonyd2wild); free-prose decode, single stream. RigBase read the GitHub README on 2026-10-28 and checked each figure against the text. RigBase did not re-run it. Speculative decoding (MTP/DFlash/DSpark) inflates predictable text; these rows prefer prose/code over counting peaks. Recipe license: MIT.
Source: https://github.com/tonyd2wild/Qwen3.8-27B-DFLASH2-AutoRound-W4A16-2x3090
https://github.com/tonyd2wild/Qwen3.8-27B-DFLASH2-AutoRound-W4A16-2x3090/blob/main/README.md
https://github.com/tonyd2wild/Qwen3.8-27B-DFLASH2-AutoRound-W4A16-2x3090/blob/main/README.md
Blank fields = not stated in the source. We never estimate build data; GPU street prices above are separately sourced and dated.
Rent before you buy: try this model on a cloud GPU for a few dollars first: Runpod (referral link; sign up with Google to get the new-user bonus)
Quick answers: Can 2x RTX 3090 run Qwen3.8 27B?