quick question: higher fps, less substeps = same simulation time?

taiman89
Posts: 13
Joined: Fri Sep 02, 2011 12:05 am

quick question: higher fps, less substeps = same simulation time?

Postby taiman89 » Tue Sep 20, 2011 3:02 am

Hey guys,

Quick question about realflow simulations. If I increase the fps of the project by say 3x and decrease min/max substeps by the same factor, will my simulation time remain pretty much the same?

And, I'm currently having to increase my substeps to about 60/120 because my particles are exploding (I think because there are too many particles being pushed too close together...), so if I increase FPS and decrease substeps by the same factor, will I start to have the exploding particles issue?


lukeiamyourfather
Posts: 2880
Joined: Mon Oct 15, 2007 4:09 pm
Contact:

quick question: higher fps, less substeps = same simulation time?

Postby lukeiamyourfather » Tue Sep 20, 2011 3:52 am

Yes, the simulation time will be about the same. Maybe a little longer since it has to write out more information to the disk. More frames equals more data to write out even though the processing is about the same.

If you're having issues because of insufficient substeps then reducing the substeps and increasing the frame rate proportionally won't fix it. You'd want to try say three times the frame rate but only half the substeps instead of a third.

taiman89
Posts: 13
Joined: Fri Sep 02, 2011 12:05 am

quick question: higher fps, less substeps = same simulation time?

Postby taiman89 » Tue Sep 20, 2011 5:18 am

Ah alright, thanks a lot for the quick response! I'd just like to increase the FPS so i can render out a slow-mo version, but I was hoping that it wouldn't take much longer than it already takes to sim it out (though I realize rendering will be much longer). Unfortunately I already have to sim at 60/120, maybe I can optimize that down a bit, but it's taking me like over 24 hours to sim a couple hundred frames (does morph daemon usually take this long? I'm using a pretty simple bird geometry...), and I didn't want it to take like 4 days for a slow motion version! Haha.

User avatar
jalixko
Posts: 72
Joined: Tue Jan 02, 2007 1:11 pm

quick question: higher fps, less substeps = same simulation time?

Postby jalixko » Tue Sep 20, 2011 10:03 am

As far as I know the Min sub-steps control the accuracy of the simulation, increasing this factor you could control particles explodes.
If you are simulating with different fps a good formula to get the correct Max sub-steps is 10.000/fps, for instance:
30fps -> 333
25fps -> 400
500fps -> 20

taiman89
Posts: 13
Joined: Fri Sep 02, 2011 12:05 am

quick question: higher fps, less substeps = same simulation time?

Postby taiman89 » Wed Sep 21, 2011 5:00 pm

Ok, thanks Jalixko!


Return to “RF5: Fluid simulation”

Who is online

Users browsing this forum: No registered users and 2 guests