From MythTV Official Wiki
Jump to: navigation, search

Wouldn't it be great if you could finish your transcoding, encoding, etc... jobs faster? It seems like that might be possible, by using some of the unused cycles of your other computers. It should be possible for mythtv users to setup a local BOINC server VM on their mythtv backends, which can distribute this work out to other local client computers for faster processing. If you are tired of waiting on h264 conversions, or whatever your particular jobs are, this may be a way to lessen that time. I will upload some public domain scripts I have modified and written to accomplish part of this task. I will also upload some notes to help you configure your machines. I have been working on configuring debian BOINC clients to accomplish the distributed work, but the images are too large to make this as practical as it could. The stuff I am posting now is more a proof of the overall concept than anything else, there is a ton of work to make this thing practical.

Location of current scripts

So, what is needed? People with more knowledge of mythtv, bash scripting, video processing, BOINC server setup, efficient linux distro rolling, and probably a lot of other stuff I haven't thought about. So if you can be of help, please contribute. Not only could this help process our video faster, but promote grid computing being on all of our machines if it isn't already. You know so we can use those extra cycles for what's really important like finding aliens.

To add more specifics about the overall operational concept of this thing... First, split up the large video file into smaller pieces. Second, use a local BOINC server VM on the mythTV backend to distribute the pieces out to BOINC clients that run client VMs to encode the video. Get the encoded pieces of video back to the backend and recombine them into the encoded video. Then you have you encoded video in less time than encoding on just the backend alone, assuming certain conditions being favorable, like the number of computers you have, the quality of the file you start with, etc...