NVIDIA A100 Tensor Core GPU, PCIe Gen 4, Passive Cooling
NVIDIA Ampere architecture is the third generation GPU architecture from NVIDIA. The A100 delivers at least 6x acceleration compared to previous generation V100’s. Designed for elastic computing, NVIDIA’s new Multi-instance GPU feature allows you to slice each A100 into 7 distinct GPU instances at the hardware level. These can be operated independently or combined as required.
- GPU Architecture: NVIDIA Ampere architecture
- GPU Memory: 40GB hBM2
- Memory Bandwidth: 1.6 TB/sec
- Interconnect: PCIe Gen 4 – 64 GB/sec. Also available with NVLink, 600GB/sec.
- Max Power Consumption: 250 W
- Thermal Solution: Passive
Read more about the NVIDIA A100 at the main XENON website.
Build Your Own Supercomputer
The A100 PCIe is now shipping. Contact Us for more details and to build your A100 based solution.
XENON also has special server builds designed for the A100 – See the details on the new XENON NITRO GX29A and XENON NITRO GX49A on the XENON website. These servers a built around PCIe Gen4 to make the most of your A100’s.
Download the full NVIDIA A100 Tensor Core GPU datasheet.
This is not a regularly stocked item.
You will be advised of delivery time-frames and asked to confirm your order, before your payment is processed. If you choose to cancel the order at that time, you will not be charged.