We take cryptocurrency — so the coins you hold can buy real hardware.Pay in crypto, spend it on real hardware. Free insured shipping worldwide over $399. Plain, unbranded boxes, sent from Germany.Free insured shipping over $399. Your rate is locked for 30 minutes once checkout opens.Rate locked 30 minutes. Every item sealed, serial-checked and covered by a 24-month warranty.Sealed, 24-month warranty.

Saved items Sign in
Dual-slot full-height card with two GPUs and 8GB GDDR5 memory

NVIDIA

NVIDIA Tesla K10 8GB GDDR5 Dual GPU Workstation Graphics Card

SKU CH-CBZ7-MK04R

Dual Kepler GK104 GPUs with 8 GB GDDR5 and 320 GB/s bandwidth deliver 4.58 TFLOPS single-precision throughput for scientific compute

  • Memory TypeGDDR5
  • Dual GK104 GPUs double throughput for single-precision workloads
  • 320 GB/sec memory bandwidth feeds compute-intensive applications
  • 58 teraflops peak single-precision performance accelerates scientific models
  • 2x 1536 CUDA cores enable massive parallel processing
  • Dual-slot full-height form factor fits standard workstation chassis

Dual-GPU compute card

The NVIDIA Tesla K10 is a dual-GPU accelerator built on the Kepler architecture with two GK104 processors. It provides 8GB of GDDR5 memory and a PCIe x16 interface in a dual-slot, full-height form factor. The board delivers 4.58 teraflops of peak single-precision performance and 320 GB/sec of aggregate memory bandwidth.

Suited for single-precision throughput

This card suits users running scientific models, seismic processing, signal analysis, image processing or video analytics that benefit from high single-precision throughput. It is not for workloads that require double-precision performance, ECC memory or display outputs. Buyers needing modern graphics APIs or higher memory capacity per GPU should consider other options.

8GB aggregate GDDR5 memory

The single specification that decides fit is the 8GB aggregate GDDR5 memory split across two GPUs, giving 4GB per GK104. Each GPU offers 1536 CUDA cores and 320 GB/sec total bandwidth. If your dataset exceeds 4GB per GPU or demands double-precision math, this accelerator will not meet the requirement.

Highlights

  • Dual GK104 GPUs double throughput for single-precision workloads
  • 320 GB/sec memory bandwidth feeds compute-intensive applications
  • 58 teraflops peak single-precision performance accelerates scientific models
  • 2x 1536 CUDA cores enable massive parallel processing
  • Dual-slot full-height form factor fits standard workstation chassis

Specifications

Memory TypeGDDR5

Questions about this item

Will this dual-slot full-height card fit in a standard PCIe x16 slot?

The card uses a dual-slot full-height form factor and connects via a PCIe x16 bus interface, so it requires a full-height case slot and a motherboard with an available x16 lane.

What is the correct way to seat the board and attach power so both GK104 GPUs initialise?

Insert the card firmly into the PCIe x16 slot until the retention bracket clicks, then connect the required auxiliary power cables from the PSU to each GPU power connector on the board.

When should I replace the thermal paste on the two GK104 chips?

Reapply thermal compound on both GK104 GPUs if temperatures rise under load or after several years of continuous operation in a workstation environment.

Also in this aisle

People compared these

Same shelf, same checkout — eight coins and a 30-minute rate lock.