请问有特斯拉板上视频编码器单元的信息吗？

目前，NVENC没有CPU回退。
这里提到了编码/解码引擎：https：//developer.nvidia.com/video-encode-decode-gpu-support-matrix
此外，您还需要特定芯片的性能概述。
因此，您需要下载包含PDF文档的Video Codec SDK。
https://developer.nvidia.com/nvidia-video-codec-sdk



以下为原文

Currently there is no CPU fallback for NVENC.
The encoding/decoding engines are mentioned here: https://developer.nvidia.com/video-encode-decode-gpu-support-matrix

In addition you need the performance overview for the specific chips.
Therefore you need to download the Video Codec SDK which contains the PDF documentation.

https://developer.nvidia.com/nvidia-video-codec-sdk

谢谢。
性能概述的分辨率为1080p。
我可以从两个1440p屏幕的编码中得到什么？



以下为原文

Thanks. The performance overview is for a resolution of 1080p. What can I expect from a encoding of two 1440p screens?

计算？
当前数字除以4应该接近。
但是你真的希望在2个屏幕上拥有30fps的相同内容吗？



以下为原文

Calculate? Current numbers divided by 4 should come close. But do you really expect to have same content with 30fps for example on 2 screens?

因此，如果编码器可以达到约600fps @ 1080p，那么可能的NVENC流数量是这个数字除以此分辨率下的平均帧速率？
如果最大帧速率超过，则所有会话/流的帧速率会下降？



以下为原文

So if a encoder can achieve about 600fps@1080p, the number of possible NVENC streams is this number divided by the average frame rate at this resolution?

And if the maximum frame rate exceeds, the frame rate drops for all session/streams?