← All systems

Reconditioned

Dell C4130

Eight Kepler GPUs across four K80 cards in a single rack unit, just under 20,000 CUDA cores addressable from one host. Built for groups that want cluster-class parallelism without cluster-class procurement.

$1,998 for the configuration below

≈20,000

Total CUDA cores

2,496

Cores per GPU

Click to enlarge
The C4130 rack unit running on a bench under a monitor and keyboard, the screen showing nvidia-smi listing eight Tesla K80 GPUs beside a Mandelbrot set render

The C4130 rack unit running on a bench under a monitor and keyboard, the screen showing nvidia-smi listing eight Tesla K80 GPUs beside a Mandelbrot set render

Click to enlarge
Dell PowerEdge C4130 opened from above: two CPU heatsinks flanked by memory slots, with four NVIDIA Tesla K80 cards along the front of the chassis

Dell PowerEdge C4130 opened from above: two CPU heatsinks flanked by memory slots, with four NVIDIA Tesla K80 cards along the front of the chassis

Hardware

Chassis
Dell PowerEdge C4130, 1U
GPU
4 × NVIDIA Tesla K80 (8 GK210 Kepler GPUs)
CPU
Dual Intel Xeon E5-2680 v4, 28 cores, 2.4 GHz (3.3 GHz turbo)
Memory
128 GB DDR4 ECC, configurable
Storage
500 GB HDD
Power supply
1400 W
GPU cooling
Forced-air; chassis rated for four K80 cards
Condition
Reconditioned, bench-tested

Preinstalled software

Ubuntu 20.04 LTS Cinnamon CUDA 11.4.48 Driver 470.256.02 GCC 9.4.0 CMake 3.31 BLAS / LAPACK NumPy / SciPy VTK

Deployment notes

A 1U server at load is loud, and its 1400 W supply draws accordingly, so plan for a rack or machine room rather than an office. Available for pickup in Boston, MA, or crated for freight.

Warranty: 1 month full warranty. The unit will be repaired at no cost.

Tell us what you need to run.

Send your workload and we'll tell you whether this machine fits.