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
Full-length dual-slot graphics card with passive aluminium heatsink and single 8-pin power connector

NVIDIA

NVIDIA Tesla M40 24GB graphics card

SKU CH-HR8K-GZK64

A 24 GB GDDR5 compute GPU with 3072 CUDA cores and 288 Gb/s bandwidth for machine-learning workloads

  • Memory InterfacePCI Express 3.0 x16
  • Memory TypeGDDR5
  • 24GB GDDR5 memory handles large datasets
  • 3072 CUDA cores accelerate parallel compute workloads
  • 288 Gb/s bandwidth feeds data-intensive applications
  • Passive cooling suits dense server deployments
  • 250W TDP fits standard datacentre power budgets

24 GB GDDR5 for machine learning

The NVIDIA Tesla M40 is a compute-focused GPU built on the GM200 architecture with 3072 CUDA cores and a boost clock near 1140 MHz. Its 24 GB frame buffer targets training and inference workloads that need large model storage. The card draws 250 W and uses a passive heatsink for dense server racks.

PCI Express 3.0 x16 bandwidth ceiling

The board connects through a PCI Express 3.0 x16 slot, limiting host transfer rates to the 3.0 generation signalling speed. An 8-pin CPU power connector supplies the 250 W TDP. Passive cooling requires chassis airflow that meets the thermal design envelope of the server platform.

384-bit memory bus and 288 GB/s throughput

A 384-bit GDDR5 interface running at 6 GHz delivers 288 GB/s of bandwidth, feeding the 3072 cores for FP32 work rated at 7 TFLOPS. Double precision FP64 reaches 0.21 TFLOPS at a 1/32 ratio. The 24 GB capacity handles large batches, while the 28 nm process and 8 B transistor budget set the absolute compute ceiling.

Highlights

  • 24GB GDDR5 memory handles large datasets
  • 3072 CUDA cores accelerate parallel compute workloads
  • 288 Gb/s bandwidth feeds data-intensive applications
  • Passive cooling suits dense server deployments
  • 250W TDP fits standard datacentre power budgets

Specifications

BrandNvidia
ModelNvidia Tesla M40
Memory InterfacePCI Express 3.0 x16
Memory TypeGDDR5

Questions about this item

Which memory capacity should I choose for this Tesla M40?

The card ships with 24GB of GDDR5 on a 384-bit bus, so there is no alternative capacity to select.

What does the 7 TFLOPS FP32 figure mean when the card is installed?

It represents the peak single-precision compute throughput of the 3072 CUDA cores running at up to 1140 MHz boost clock.

How does the PCI Express 3.0 x16 interface limit performance?

The x16 link provides up to 16 GT/s per lane, which is the maximum host bandwidth the GM200 GPU can use.

Also in this aisle

People compared these

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