This is impossible to demonstrate without sending a massive patch, as it seems to be related to patch size.
I have a couple of big patches at the moment, and in certain parts of the patch (not specifically related to big spread counts) when I link a pin to start a link I can have to wait for several seconds before being able to complete the link, if I right click to do multiple, after the pause I can do them as normal, its only starting the link that causes the issue.
This patch has multiple copies (3-6) of a patch that has 40 odd sub patches in it + other stuff, I’m using evaluate to turn off whats not being used at anyone time.
I’ve tried disabling undo but the problem persists, the memory used by the patch is 700MB+ so I can imagine that it could be related to the footprint.
Anyone else seen this behaviour, or have any suggestions?
i also had this once in a project. it seems to be related to vvvv’s checking which links would cause a feedback… but it would need such a massive patch to profile it and find the exact cause.
if you feel bored you could delete some links which connect big parts of the patch and see which ones stop vvvv from looking too deep for cyclic graphs.
it would mostly be interesting how many nodes this patch has itself (and then how many pins those nodes have alltogether). from what you are saying it must be a massive patch with tons of nodes. nodes in subpatches should not contribute to such patching lag, on the contrary: if you get such a lag this is a hint your patch is too big and you should do more subpatches.
Each subpatch is a reasonable size, but they nest fairly deeply, say 4 or 5 levels.
Its not always the patch with many subpatches in, or many nodes where the slow down happens, which is what I found odd, there is a control loop however that has 510 slices and gets processed by the interface, modified by midi, saved and loaded as preset values, and then loops round again, but 510 slices isnt that big either.
To much on for deep checking at the moment, but maybe next week I’ll have a look!
Ah, I just deleted the final link before the values loop around and that stops it, so it must have to search through all the patch looking for loopbacks before allowing the link, I can maybe add a switch to let me patch more swiftly for now, and next week try and optimise!
well sounds interesting that you obviously found the reason. i don’t get it though from your description. would be great if you could boil this down to a simple setup that demonstrates the issue and upload here.