Lewis Taylor

You don't want to be downgrading the amount of ram you have.
Taking advantage of different acceleration is meaningless if the data can fit in ram.
It really feels like 24gb is the main baseline for GPU compute in Houdini, anything less
tends to limit the amount of complexity to an annoying degree.

L

There are a few reasons, but like I said, it's generally best to learn the workflow of not
kicking things off that are wrong/you want to interrupt.
I know it's not answering your question, but it's a known issue on heavy sims, due to all the parms that are being initialized and queried, so we tend to direct Artist's to be a bit
more mindful before launching a sim.

Escaping renders crashes houdini, crashes Maya, crashes Max. Interrupts aren't trivial.

L

Yes, and in those nodes the overhead is very low, or acceptable.
I was more referring to your example, stashing a heavily subD'd bit of geometry.


L