Lot of 4x NVIDIA
Tesla Referral K40c 12GB GDDR5 Computing Processors
This listing is for a bulk lot of four (4) NVIDIA Tesla Referral K40c active-cooled workstation GPUs, providing a combined 48GB of GDDR5 video memory. Based on the NVIDIA Kepler architecture (GK110B), these units are HP OEM validated (Part Number F1R08A) and are designed specifically for high-performance computing, scientific simulation, and AI workloads. These units were pulled from a working enterprise environment and have been fully tested for performance and stability.
Tested & Working
Each unit in this lot has been rigorously tested and is in Grade B condition. You can expect light cosmetic wear consistent with professional use in a server or workstation environment. There are no damaged components, and all PCI Express contacts are clean and intact.
Key Features & Specifications
- Quantity: 4 Units (Total 48GB VRAM)
- Brand: NVIDIA / HP OEM
- Model: Tesla Referral K40c
- HP Part Number: F1R08A
- MPN: 900-22081-0340-000
- VRAM: 12GB GDDR5 (Per Card)
- CUDA Cores: 2880 (Per Card)
- Compute Capability: 3.5
- Interface: PCIe 3.0 x16
- Form Factor: Dual Slot (Active Cooling)
- TDP: 235W (Requires external power)
- Architecture: Kepler GK110B
What's Included
This listing includes four (4) bare NVIDIA Tesla Referral K40c GPUs only. Heatsinks are integrated (Active Cooling), but power cables, original retail packaging, and manual/drivers are NOT included.
Compatible Systems Enterprise Workstations: HP Z-Series (Z440, Z640, Z840), Dell Precision (T5810, T7810, T7910), Lenovo ThinkStation P-Series.
Server Compatibility: Compatible with any server or desktop featuring a PCIe 3.0 x16 slot and sufficient PSU headroom (235W per card).
Note: Ensure your power supply has the necessary 6-pin and 8-pin PCIe power connectors to support these units. Ideal For
- AI & LLM Development: With a combined 48GB of VRAM, this lot is excellent for running smaller Large Language Models (LLMs) such as Llama 2 (7B/13B quantized), Mistral, or BERT models where VRAM capacity is the primary bottleneck.
- Homelab & Parallel Computing: Ideal for CUDA programming, GPGPU development, and scientific research using Amber, GROMACS, or LAMMPS.
- Video Rendering: Offloading compute-heavy tasks in supported professional suites.
NVIDIA Tesla Referral K40c, K40, 900-22081-0340-000, F1R08A, Kepler GPU, GDDR5 12GB, 48GB VRAM Lot, CUDA Compute, Deep Learning GPU, LLM Inference, HP Z840 GPU, AI Training, GPGPU, PCIe 3.0 x16, 2880 CUDA Cores, Tesla Referral Workstation Card.
Important Information All items are pulled from working enterprise environments and inspected by our team. These units show minor cosmetic wear. No manufacturer warranty is included. Buyers are responsible for verifying compatibility with their specific motherboard and power supply. Sold as-pictured without additional accessories. Returns & Questions
Returns accepted only if the item received does not match the listing description. If you have any compatibility or condition questions, please message us before purchasing and we will respond promptly.
