Mark Tucker

We are not planning on tackling this in the upcoming release cycle.

I would strongly recommend figuring out a way to have your camera prims created at the top of your graph.

Why are your camera nodes time dependent? Can you use a Cache LOP to eliminate the time dependency?

The reason that python code won't work is that there is no dependency set up. It may work once, but when you change renderers, there is nothing to notify the Switch LOP that it needs to recook.

There is an outstanding RFE for allowing a python script to run when switching delegates. This mechanism could be used to drive a global context option value, and you could switch based on the context option value. Not sure if/when this RFE will be implemented, but this is the most likely path forward on this request (assuming it's not just a matter of different render settings as suggested by @tamte).

The most common thing I've seen leading to this kind of situation is (sometimes unintentional) caching within the LOP network. As an example, we recently got a bug submission where the Motion Blur LOP was causing memory bloat like this. It was an incorrect setting on the Cache LOP inside that HDA, which has now been fixed. So maybe look for Cache LOPs and Cache SOPs (especially ones buried inside HDAs).

Another interesting data point would be to know if you can reproduce without using Karma. This would help localize the problem to either being Karma causing the bloat or LOPs causing the bloat.