Live
AI hardware

NVIDIA A40 PCIe

FP16 throughput
1.5×10¹⁴ FLOP/s
Memory
44.7 GB
TDP
300 W

NVIDIA A40 PCIe is an AI accelerator designed by NVIDIA (GPU), released in October 2020. It is built on a 8 nm process at Samsung, packing 28.3B transistors.

It delivers 1.5×10¹⁴ FLOP/s of tensor FP16/BF16 throughput. It carries 44.7 GB of memory at 648 GB/s of bandwidth. Its thermal design power is 300 W.

Full record
Manufacturer
NVIDIA
Type
GPU
Release date
Oct 5, 2020
Tensor FP16/BF16
1.5×10¹⁴ FLOP/s
FP32
3.7×10¹³ FLOP/s
FP64
5.8×10¹¹ FLOP/s
INT8
3×10¹⁴ OP/s
Memory
44.7 GB
Memory bandwidth
648 GB
TDP
300 W
Process node
8 nm
Transistors (millions)
28,300
Die size (mm2)
628
Foundry
Samsung
Tensor cores
336
More from NVIDIA
SourceEpoch AI, 'Machine Learning Hardware'. Published online at epoch.ai. Retrieved 2026-07-29 from https://epoch.ai/data/machine-learning-hardware. Licensed under CC BY 4.0.
← All ai hardware