Multi computer processing via (w)lan

I would like to see an option to let several seperate computers connected via WLAN for example to work on the same video file via Topaz Video Enhance AI. Network transfer is quick nowadays and this could shorten the processing time a lot. :slight_smile:

:roll_eyes: :laughing: :woozy_face:

??? not understand…

Basically the questions is whether networked computing could be possible.

Scenario: Computer A is the main computer with Topaz Video AI, then computer B in the same network is running topaz CPU/GPU NODE, which receives computational instructions from computer A and thus adds to the overall computation capacity of the Video AI.

I just cut up the video in multiple sections and render them on multiple Mac Mini M1s. Then reassemble in the end in FCPX. I usually want to tweak the video quite a bit afterwards anyway, so its not too tedious. Looking at finding a Script to do this work for me however.

Two hiccups:

  • Gigabit networking is not really sufficient for high-bandwith video.
  • a 2nd high-end video card to make use of mGPU models is going to be a much easier and productive option that avoids the network bottleneck

So unless we’re talking several fully-stacked server-grade devices running on Infiniband I can’t see this being efficient.

Just update my configuration, I use a 2.5G networking setup (the adaptors are so cheap now) and wrote a script to take a MP4 file, split it up and then feed it to my 8 Mac Mini M1 computers. They then script out to run the software, render and then hopefully finish. I then have a script to recombine all of the files back into one.

It’s not fully automatic and fool proof - I wish the software had a command line interface, but I’m not aware of anything like that available.

But all in all it works just well. A one hour video that did take .10 fps to render now is completed in about 23 minutes total (if I did the math right - add another 5 to 10 minutes for editing the final file for cleanup of borders and tweaking color and sound). This is providing it a 480p file, rendering with Proteus and outputting a 720p file.