AWS Thinkbox Discussion Forums

Houdini Frame by Frame caching (parallel caching)

Hello,

I’m working on a simulation which includes FLIP and RBD. I don’t really need/want to cache the .sim files, only need the results of the DOP network, but I’m having some troubles when I try to cache both outputs in parallel.

Let me try to explain: I’m using two Geometry nodes in the out context. Each one point to a different object, in my particular case one is for the compresses fluid cache and the other for the RBD cache, so each one is pointing to a totally different place in the obj context. Then, in out context, I use a merge node set to Frame by Frame, and hit render. This allows “parallel caching” so the sim is calculated one time but each frame but I can write the different outputs to disk (flip.001, rbd.001, flip.002, rbd.002, and so on)

I’m trying to send this merge to deadline but the result is two different jobs, one for the flip cache, and one for the rbd cache, which of course would mean simulating two times (one for the first job, and when in finished, simulate again to cache the second job). This is no good at all because the sim takes around 10 hours to finish…

Is there a way to send those caches as parallel jobs as it can be done from within the interface? In other words, one job, two different outputs.

Simulating from the interface is not a good thing to do because the memory consumption is way higher (just because the scene is opened) and also is somehow unstable… I have noticed it simulates some frames but then it just hangs (not always the same frame, and not always happens, but this inconsistency is what I’m trying to avoid). Simulating in deadline works like a charm, but the issue is I need to simulate twice which is no good either.

I hope there’s a hidden option somewhere that I missed…

Thanks in advance for any help,
Orlando.

Privacy | Site terms | Cookie preferences