Is Ampere architecture compatible with cuDNN 7.6.3 and CUDA 10.0.130

Accelerated Computing CUDA CUDA Setup and Installation
1

Hi NVIDIA,

We are planning to purchase the A10 or A100 GPU card and we use the PyTorch library which uses the 7.6.3 version of cuDNN and 10.0.130 CUDA.
According to cuDNN Support Matrix :: Deep Learning SDK Documentation, I don’t see Ampere GPU architecture is part of it. Could I confirm the 7.6.3 cuDNN doesn’t support Ampere GPU architecture?

Looking forward to your reply.

Thanks,
Tim