Buying in volume? Request a bulk quote
We source this item through our supplier network and do not hold inventory. Price and availability are confirmed with you before any order is accepted.
Questions about this part? Talk to us.
Phone +1 872 293 4196 · Email supply@allchipsupply.com
Location: USA · Shipping: Worldwide Shipping
The NVIDIA A100 40GB PCIe is the 40 GB version of NVIDIA's Ampere-generation A100, still widely deployed for AI inference, training and HPC. It carries 40 GB of HBM2 memory with 1,555 GB/s of bandwidth and delivers 19.5 TFLOPS of FP64 Tensor Core performance, according to NVIDIA's datasheet.
The card connects over PCIe 4.0 x16 and has a maximum power of 250 W. Multi-Instance GPU can split it into up to seven isolated instances of 5 GB each, which is useful for shared inference servers. It is a dual-slot, full-height, full-length passive card, so it needs a server with directed airflow. Check your server's qualified GPU list before ordering.
| Architecture | NVIDIA Ampere |
|---|---|
| GPU memory | 40 GB HBM2 |
| Memory bandwidth | 1,555 GB/s |
| Interface | PCIe 4.0 x16 |
| Form factor | Dual-slot, full-height full-length, passive |
| Max power | 250 W |
| Multi-Instance GPU | Up to 7 MIG instances @ 5 GB |
| FP64 Tensor Core | 19.5 TFLOPS |
Specifications are summarized from manufacturer and listing information; confirm against the manufacturer datasheet before purchase.
Manufacturer datasheet / product page (opens the manufacturer's website)
NVIDIA Ampere · 80 GB HBM2e, ECC · ~1.9 TB/s
NVIDIA Ada Lovelace · 18,176 · 48 GB GDDR6 with ECC
NVIDIA Hopper · 80 GB HBM2e · 2.0 TB/s
NVIDIA Hopper · 141 GB HBM3e · 4.8 TB/s
