Training-class accelerator with 1.5 TB/s of memory bandwidth.
Multi-Instance GPU partitions one card into up to seven isolated instances.
250 W passively cooled — requires a chassis rated for GPU airflow, such as a Supermicro or Dell GPU platform.
Used for model training, HPC and large-scale analytics.
Specifications
Brand
NVIDIA
Model
A100 PCIe
Type
Datacenter GPU
Capacity
40 GB HBM2
Interface
PCIe 4.0 x16
Form Factor
Dual-slot, full height
Compatibility
GPU-capable rack servers
NVIDIA A100 40GB Tensor Core GPU is supplied by Interseller, the enterprise
hardware store of QuantaRock Systems. Pricing is quoted on request —
contact us for a quotation.