unstable initial fill problems and liquid smoothing

[squintnic]

here is my current scene file, in case anyone can shed any light on the intial fill wobbling or ideas for foam control / previewing
i feel like by upping the scene scale i could get the fluid to settle but then the splash scale would feel wrong, currently feels a bit out of whack with the size of the glass / strawberry

https://www.dropbox.com/s/3ojbph83tf...st_07.max?dl=0

[Svetlin.Nikolov]

Hey,

25 frames might indeed be insufficient for the liquid to settle - I'm afraid you need a longer pre-roll, or maybe you could use a weak Drag force on the liquid?

If the CPU is dual, it's possible that the simulation would run even faster on just a single node - do you get a choice for multiple NUMA nodes so you can exclude all but one?

Maybe the Liquid resimulation would work for you? It allows you to keep the liquid shape as it is in the original simulation, but to change the Foam settings over the existing liquid.

[LarsSonparsson]

Just a quick question. How do you exclude a NUMA node?

[Svetlin.Nikolov]

Ah - there's a list in the Simulation rollout which gets populated with the nodes, if they are available.

[LarsSonparsson]

Yes, I have two nodes there, but I don't know how to disable the other one. They are selected (blue), non selected (grey), but there is no other options or check box or something...
Sorry for stealing this thread for my own question...

[Svetlin.Nikolov]

Only the selected ones will be used for simulating - I think ctrl+click would toggle a node

[LarsSonparsson]

Ah, exactly, just got it working. Thanks!

[squintnic]

ah thanks i didn't know that NUMA trick either - will give that a try along with a longer pre-roll and liquid re-sim. thanks for the assistance guys im getting there

(Im looking at a new i9 chip especially for simulating would fewer cores at higher clock speed be better foe phoenix fd?)

[Svetlin.Nikolov]

Higher clock would be better. Check the sticky topics on recommended hardware