What Nvidia GPUs can I use for TensorRT execution provider in ONNX runtime?

[ExecutionProvider]

Not all Execution Providers (EP) offer the same acceleration platforms for ONNX Runtime (ORT) as for the corresponding inference engine. For instance, the Vitis AI EP only offers support for their Alveo U200 and U250 acceleration cards.

I want to know what Nvidia GPUs I can use if I choose to go with the TensorRT EP and ORT. Also, do they need CUDA support?

This is all the information the official docs provide:

The TensorRT execution provider in the ONNX Runtime makes use of
NVIDIA’s TensorRT Deep Learning inferencing engine to accelerate ONNX
model in their family of GPUs.

The TensorRT inference engine lists the supported hardware here, but I am wondering if this list is valid for the TensorRT execution provider as well.