the official info of various models of GPU support what kind of accelerate mode

Hi, all.

I am using tensorRT to accelerate model-infer-process, but i found that different models of GPU support different mode
For example, gtx1080ti only support float32 and int8, not support float16

what about other models of GPU ? p100, p4, V100 and so on.

Where can i get the compete official information about this?

What is more, what kind of GPU support DLA? (all i know is gtx1080ti not support)

Hello,

Please see: https://docs.nvidia.com/deeplearning/sdk/tensorrt-support-matrix/index.html#hardware-precision-matrix

Thanks.