i think i am misunderstanding some fundamental points with boygrouping. help would be appreciated.
starting question to get my bearings…
- i would like to choose which nodes are used on which clients.
eg. i have kinect processed on one client, audio/beat tracking on another, server making decisions, and renders on the others. would like to keep the processing seperate and feed data between.
( can i do this? / is this the right way of thinking? )
(on a side note. i could do with stopping the windows moving around on the client when i move the window on the server. i use vnc with combos of single and dual-screen, so the windows keep going off the edge.)
ClientID is what you need to use
ja, seems so. boygrouping is not for distributing computation and sending results to another pc for rendering.
boygrouping is a simple server/client render cluster setup. you don’t touch the clients at all (apart from starting them up so they connect to the server). from that on they just render what the server tells them to. windows are showing on clients only for debugging reasons (which ideally should not be needed).
of course you can still distribute processing of your inputs to different pcs and send back data to a server but then boygrouping will not help you. for such you’d use the tcp or udp nodes…
does that help your understanding?
that helps. i have my server and renderers boygrouped, and the tracking/audio seperate (but still feed data through network).