Headless Ubuntu box with RTX 3090 serving code completion to laptops
A developer runs Ollama on a headless Ubuntu desktop with an RTX 3090 and connects laptops via the Continue VS Code plugin. The 32B model was too heavy in practice, so 7B is used.
Solo2-10 peopleCoding$1k-3kSensitivity: High (client/confidential)
Hardware
| Item | Approx. price (as of 2026-10-16) |
|---|---|
| Desktop PC with NVIDIA GeForce RTX 3090 (24GB) | Not stated in source |
Source prices are as stated by the source. Retailer offers: prices checked 2026-10-16, may change. Some links may be affiliate links; we may earn a commission at no extra cost to you. This never changes what we list.
Software stack
- Ubuntu (headless)
- Ollama
- Continue (VS Code)
Models used
- qwen2.5-coder:7b (chat, edit, autocomplete)
- Tried Qwen2.5-Coder-32B (~21GB VRAM) - too big for interactive use
Good at
- Author 'satisfied with the quality'; no noticeable sluggishness with 7B
- Works like Copilot, no adaptation needed
Weak at
- 32B model sometimes froze the Continue plugin
Rough total cost
Not stated in source.
Getting started
- Install Ollama on the GPU box per official docs.
- ollama run qwen2.5-coder:7b to pull and test.
- Install the Continue extension in VS Code.
- In Continue config.yaml set provider: ollama, model qwen2.5-coder:7b, apiBase pointing at the server.
- Assign roles chat/edit/apply/autocomplete.
Source & credit
Local coding assistant with Ollama and VS Code Continue plugin — Arul Selvam Periyasamy (blog)
Summarized from the source; quotes are the source author's words. Check the original before buying.