Two 24 GB cards give 48 GB total; runtimes such as llama.cpp can split a model across GPUs. Check slot spacing, PSU capacity and case airflow.
| Part | Choice | Price / what to check |
|---|---|---|
| GPU | GeForce RTX 3090 (24 GB) | $1,499 launch MSRP |
| GPU | GeForce RTX 4090 (24 GB) | $1,599 launch MSRP |
| CPU | Current 6–8 core desktop CPU | Any recent 6–8 core chip is enough; LLM speed depends on the GPU |
| Motherboard | Matching socket, two x16-size slots | Check slot spacing for two cards and x8/x8 lane support |
| RAM | 64 GB DDR5 | Two sticks for dual-channel; see RAM prices |
| Storage | 1–2 TB NVMe SSD | A Q4 70B model file is ≈40 GB; leave room for several models |
| PSU | Per GPU maker's recommendation | Use the wattage on the GPU maker's spec page |
| Case | Fits GPU length/thickness | Check the case's max GPU length against the card |
Only GPU prices are cited (NVIDIA launch MSRP). Other parts are listed with what to check rather than a price, because we do not yet track live prices; the total may differ from the headline budget. Target example: Llama 3.3 70B at Q4 needs ≈47 GB.
| GPU | VRAM | Memory bandwidth | Launch MSRP (US) | Source |
|---|---|---|---|---|
| GeForce RTX 3090 | 24 GB | 936 GB/s | $1,499 Launch MSRP (2020); sold used today. Used price TBD. | NVIDIA: GeForce RTX 3090 specs; NVIDIA Newsroom (Sep 2020): RTX 30 Series pricing; Wikipedia: GeForce RTX 30 series (memory bandwidth table) |
| GeForce RTX 4090 | 24 GB | 1008 GB/s | $1,599 Launch MSRP (2022). Current price TBD. | NVIDIA: GeForce RTX 4090 specs; NVIDIA Newsroom (Sep 2022): RTX 40 Series pricing; Wikipedia: GeForce RTX 40 series (memory bandwidth table) |
VRAM and bus width from NVIDIA spec pages; bandwidth from Wikipedia series tables (not listed on NVIDIA pages); MSRPs from NVIDIA press releases. Current street prices: TBD.
Try other budgets
AI PC Builder
Loading…