Forums Search
Found 183 posts.

Search results Show results as topic list.

Houdini Lounge » problems with higher res FLIP sims

User Avatar
johner
790 posts
Offline
 
JordanWalsh
cool, thanks for the tips!
So is the cache just there for collision objects and not the actual sim?

No, it's there for the sim, and very helpful with medium-sized simulations. But past a certain size simulation, there's not much chance that it will all fit in memory (even on a 24Gb machine), at which point it starts paging to disk and gets pretty slow.
See full post 

Houdini Lounge » problems with higher res FLIP sims

User Avatar
johner
790 posts
Offline
 
Glad to hear things are working now.

On that note, whats the best way to use the cache amount?

I'd recommend turning off Allow Caching on the FLIP FluidObject, since then you'll never risk running out of cache due to particle count. I also usually turn off Allow Caching to Disk on the DOPNet, since 1) it can be slow once you hit the disk, and 2) you're not surprised when things work differently in the UI from hbatch or hython.

With those settings a 1Gb cache or so should usually be plenty, since you only care about caching the collision objects.
See full post 

Houdini Lounge » problems with higher res FLIP sims

User Avatar
johner
790 posts
Offline
 
What do the Cache settings on the DOPnet look like? If you're using DOPs collisions (i.e. your collision object is a Static Object), then DOPs needs the previous frame in the cache to compute good collision velocities, especially while substepping the FLIP sim.

It's possible that if you have really small cache settings, then that one frame is not fitting in memory and is being cached to disk when running in the Houdini UI. But Cache to Disk defaults to Off for non-interactive sessions like hython or hbatch, so you won't get good velocities when simulating on the farm.

Just a theory. What do those parameters on the DOPnet look like?

Edit: another thing to try is to turn off Allow Caching on the Creation tab of the FLIP Object, since it's really only important that the collision objects are cached.
See full post 

Technical Discussion » Houdini 12 gold Flip slover no particle ID

User Avatar
johner
790 posts
Offline
 

Houdini 12.1.30: The FLIP Solver has a new option to add a unique id attribute to each particle. Additionally, the Reseeding operation has been updated to properly create new ids when generating new particles.


Also in 12.0.706
See full post 

Technical Discussion » animated Fan Force on FLIP

User Avatar
johner
790 posts
Offline
 
Did you set the animated parameters on the Fan Force to “Set Always”?
See full post 

Technical Discussion » Curl Calculation

User Avatar
johner
790 posts
Offline
 
icarus551
I was shown a similar solution which is correctly theoretically, but if you compare it against the dop computed curl they don't even look similar.

I ran a 2D smoke sim (xz axis) to compare. Curl vectors of a 2d field should all point along one axis (y axis). The dop computed curl does this but none of the sop solutions do.

Hmm, I'm not seeing much discrepancy. The attached 2d smoke test creates a curl attribute from both a DOP GasAnalysis-generated field, and the SOP solution. The values aren't identical, but are very close. Switch the smoke sim to 3d and the values are still very close.
See full post 

Technical Discussion » Curl Calculation

User Avatar
johner
790 posts
Offline
 
The VolumeAnalyis SOP should really have an option to calculate the curl, but in the meantime you can compute the gradient of each velocity component and calculate the curl by hand in VOPs (thereby doing a little extra work: nine derivatives instead of the six minimally required).

See attached for an example of computing the curl of a FLIP sim (and turn on the curl attribute visualization).
See full post 

Technical Discussion » Keeping consistent behavior when increasing FLIP res?

User Avatar
johner
790 posts
Offline
 
nikmolson
I also have a problem where it seems that when i emit the fluids higher than a certain velocity, the leading parts of the fluid seem to flatten and i get kind of flat leading faces of the sim, this seems to get worse when i upres aswell.

You probably need to increase the Max Cells to Extrapolate setting under Volume Motion | Solver. With low substeps and high resolution volumes, the particles can move outside the bounds of the underlying volume in one timestep and run into zero velocities, hence the flattening. Increasing this value should alleviate the problem.
See full post 

Technical Discussion » H12 flip emitter rate?

User Avatar
johner
790 posts
Offline
 
The SourceVolume DOP has a Scale Velocity parameter. Have you tried that, or is it not doing what you want?
See full post 

Technical Discussion » H12 flip emitter rate?

User Avatar
johner
790 posts
Offline
 
Are you trying to reduce the velocity of the particles that are being emitted, or just the quantity of them?
See full post 

Technical Discussion » affecting sourced particles on H12 flip with popsolver?

User Avatar
johner
790 posts
Offline
 
What version are you using? There was a bug fixed recently that makes particles from SourceVolume DOPs work with POPSolvers:


Monday, April 30, 2012
Houdini 12.0.612: When sourcing particles, the SourceVolume DOP will now add any new particles to an already existing particle system, which allows such particles to work correctly with POPSolvers.
See full post 

Technical Discussion » FLIP FLUID Attributes

User Avatar
johner
790 posts
Offline
 
circusmonkey
As the surface falls FLIP does struggle to hold itself together.

The section on viscosity [sidefx.com] in the docs has some info that might help keeping the viscous fluid together. Some of the things that work well for inviscid fluids, like reseeding and automatic detection of ballistic particles, can be less useful or even counter-productive with highly viscous fluids. This issue can crop up as well with high viscosity values:

The FLIP Solver computes the number of substeps to calculate based on velocity. While the viscosity solver should always be stable, it will be more accurate with smaller substeps. A slow moving, highly viscous sim might need more substeps than the 1 that the FLIP Solver will give it based on its velocity. Increase Minimum Substeps until you get sufficient viscosity.

Unfortunately the viscosity solve is fairly expensive, so increasing minsubsteps can be painful but necessary sometimes.

eetu
Here's another variation, melting by transfering low viscosity by a metaball.
Cool!
See full post 

Technical Discussion » FLIP FLUID Attributes

User Avatar
johner
790 posts
Offline
 
Viscosity is specified in units of dynamic viscosity, or kg / (m·s), so it does tend to have high values. In particular it will be divided by density, so the same viscosity value will have more of an effect on less dense fluid.

If you're trying to do a variable-viscosity effect like melting, you might want to set a high viscosity value on the FLIP Object's itself via the Physical tab, then use 0-1 viscosity values on the FLIP particle attribute, with the Viscosity attribute Mix Method set to multiply.
See full post 

Technical Discussion » Water droplet on surface - flip oddness

User Avatar
johner
790 posts
Offline
 
If you watch the droplet as it falls relative to the bounding box, it's basically
“outrunning” the bounding box as it speeds up. To correct this, you need to increase the Max Cells to Extrapolate parameter on the Volume Motion | Solver tab. From the help:


Max Cells to Extrapolate

The number of non-fluid cells that should be filled with velocity values on
the non-fluid side of the velocity field. Increase this value for very fast-
moving fluids and/or a low number of substeps.


Alternatively (or additionally) you can increase the number of substeps, so the fluid doesn't move as far in each timestep.
See full post 

Houdini Lounge » Houdini Flip fluids meshing problem

User Avatar
johner
790 posts
Offline
 
There's some tips for working with viscosity here:
http://www.sidefx.com/docs/houdini12.0/dyno/liquids#viscosity [sidefx.com]

In particular, the Average Position surfacing method often leads to the artifacts you're seeing. They're less noticeable with fast moving water, but for slow, viscous fluids I usually use the Metaball method.

You probably also want to turn off reseeding, which tends to fill in the gaps between parts of the fluid surface.
See full post 

Houdini Indie and Apprentice » Floating rigid bodies on Flip fluids.. possible?

User Avatar
johner
790 posts
Offline
 
jeff
The feedback value usually has to be pretty high. Impulse force to counteract the mass of the object that you wish to float. Try values between 50-2000 if the sim fits on the default Houdini grid.

This has actually changed for H12, since the FLIP Solver now takes into account fluid density (which defaults to water = 1000) when solving for pressure, which in turn scales the feedback force. The H11 solver was equivalent to treating fluid density as 1.

From the Feedback Scale parameter doc:

A scale factor used in applying feedback forces to other objects. A value of zero prevents any feedback from occurring.

The default RBD object has the same density as water, so to balance with the default fluid density a value of 1 should be used.


Also, the “DensityViscosity” FLIP example shows a combination of variable density, variable viscosity, and a floating RBD object.
See full post 

Technical Discussion » Elastic effect on flip fluids on H12

User Avatar
johner
790 posts
Offline
 
After re-reading that paper, I updated (only) the parameter labels on the OTL above.

So:
Elastic Modulus - the strength of the elasticity effect

Plastic Yield Limit / formerly Gamma - the strain limit at which plastic flow occurs. A high value means little plastic flow (fluid is mostly elastic and always tries to return to original shape), low value means plastic flow can occur when norm of the strain matrix (think total deformation) is above this limit.

Plastic Flow Rate / formerly Alpha - how quickly the fluid will flow when strain is beyond the yield limit. A low value (or zero) will imply mostly elastic behaviour no matter the yield limit; a high value means the fluid will flow quickly once past the plastic yield limit.
See full post 

Technical Discussion » Elastic effect on flip fluids on H12

User Avatar
johner
790 posts
Offline
 
I should have mentioned, if you're wondering what those obscure Alpha and Gamma parameters represent on the GasStrainIntegrate DOP, they're from this paper (which is actually somewhat readable as these things go):

http://graphics.berkeley.edu/papers/Goktekin-AMF-2004-08/ [graphics.berkeley.edu]

The paper and video there both have some demonstrations of different parameter values.
See full post 

Technical Discussion » Elastic effect on flip fluids on H12

User Avatar
johner
790 posts
Offline
 
Nice work figuring the strain stuff out, Ian. I think you might still be missing advection of the strain matrix, and as Jeff points out, that's tricky with FLIP since ideally the particles represent the entire state of the simulation at any given time.

I put together a little FLIP elasticity asset that uses the Strain DOPs and FLIP/PIC for advecting the strain matrix, storing the strain in attributes on the particles. The substepping requirement for the explicit GasStrainIntegrate DOP makes it pretty expensive, as does GasParticleToField on a matrix attribute, but it might be useful in some contexts. Also, in my tests it seems like adding a small amount of viscosity helps damp / stabilize the elasticity force.

There may be a bug in Gas Velocity Stretch's handling of matrices, so at the moment I've got that DOP bypassed within the OTL (and offhand I'm not sure the rotation works with a FLIP-style update anyway).
Edited by - March 22, 2012 14:07:23
See full post 

Technical Discussion » Open CL Settings

User Avatar
johner
790 posts
Offline
 
There should be a significant speedup from 12-physical CPU cores to a Quadro 4000, but you probably need to increase the resolution higher than 64^3 before it's apparent.
Edited by - March 8, 2012 15:03:37
See full post 
  • Quick Links