Dell R730 AI Home Lab Server 128GB RTX 3060 Ti Dual Xeon 20-Core For Sale

Dell R730 AI Home Lab Server 128GB RTX 3060 Ti Dual Xeon 20-Core
When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.

Buy Now

Dell R730 AI Home Lab Server 128GB RTX 3060 Ti Dual Xeon 20-Core:
$999.00

Hardware:- Two Intel Xeon E5-2660 v3 processors, 20 cores / 40 threads total- 128 GB DDR4 ECC registered RAM, eight 16 GB DIMMs at 2133 MT/s- Dedicated graphics card included: NVIDIA GeForce RTX 3060 Ti, 8 GB VRAM; tested for local AI inference through Linux VM GPU passthrough- NVIDIA GeForce RTX 3060 Ti with 8 GB VRAM- Dell PERC H730 Mini- Four Intel I350 Gigabit Ethernet ports- Two installed 1100 W power supplies
Demonstrated uses:Local coding assistants, chat/summarization experiments, Linux VMs, containers,and an OKD/Kubernetes home lab. GPU passthrough was demonstrated from the nativeOKD/KubeVirt host into an Ubuntu VM with 12 vCPUs and 32 GiB RAM.Nested KVM is enabled and its API is available inside the guest; a second-levelguest boot has not been tested. No claim of GPU passthrough through two VM layers.
Measured local generation performance on this hardware:Qwen2.5-Coder 7B Q4_K_M: 75 tokens/sec median.Llama 3.2 3B Q4_K_M: 133 tokens/sec median.Qwen2.5-Coder 1.5B Q4_K_M: 200 tokens/sec median.Ollama 0.35.1, GPU-backed Linux VM, short identical prompt, 4096-token contextsetting, three warm 256-token outputs. Loading and prompt-processing time excluded.Initial model loading took approximately 14-38 seconds across four tested models.These are single-request throughput measurements, not model quality or concurrencyguarantees. Larger models/longer contexts have different memory and speed requirements.
Buy Now

Documentation