08-28-2020 04:42 PM - edited 08-28-2020 08:15 PM
Hi. I want to develop an accelerated FFmpeg plugin to be used on AWS EC2 F1.
I know the XMA doc, but there is not an example there. I saw this video (2019-03-04) but it is about SDaccel and I know that there were changes on the 2019.2 version (for example, the YAML file was deprecated). The ABR-video-transcode repository is also outdated.
Do you know an updated resource? It is ok if about Alveo instead of AWS.
09-01-2020 08:25 AM
Any idea? Maybe a thought? will be XMA discontinued and is it the reason for the outdated repositories and missing new examples? (I don't hope that but I don't understand why there is not updated information).
09-08-2020 03:42 PM
It seems that I am asking for something weird... Anyway, I think that I am misunderstanding what is solved in XMA in conjunction with FFmpeg. Can I take advantages of some already developed decoders and encoders? Is XMA providing some of them? If I develop my own accelerator, why to develop an FFmpeg plugin? Which is the improvement that I can get? Answers will be welcome.
09-08-2020 04:12 PM
This image is from Vitis Libraries 2020.1:
I know that TensorFlow and Caffe accelerated versions are provided into a docker of Vitis-AI. Where is the accelerated FFmpeg version provided?