hi,

i would like to ask experienced Houdini users about the most “houdinish” approuch in the following scenario, because I find my approach little bit ugly.

so the situation:

1) i have one DOP network(A) and also one SOP solver(B).

2) there is an interaction between them in the following way:

3) in A, i use “object merge” and import B. I am using vex/vop to find points in B and in some conditions, i copy some of them to A.

4) in B, i am deleting points that were copied to A

my approach is:
A) in A, i have some int @added_list where i store point indices of the copied points
B) in B, i read thes array, and delete the selected points

the problem is, that each solver has own timestepping and it read one frame old geometry from the second solver, so the code is very ugly to ensure that the second side has read the list so i can clear it and fill with new points.

So are there some workflow hints to do it in the corrects way?

thanks

For (B), are you're using the Solver SOP from the SOP context? If so, maybe you can try manually setting up the SOP solver inside the same DOP context. Under the hood, the Solver node in the SOP context is actually just a DOP solver, called the SOP Solver, that's wrapped up nicely.

So you could chain the solvers using the “Multiple Solver” DOP. This means within the same DOP Network, your original simulation in (A) can simulate first, then the result will be passed to the SOP solver, and be processed by your SOP nodes from (B).

Hi,

thanks for a hint.

I already tried this, but I finally abandoned this, because I had problem with “Prev_Frame” node, that is included in the “Solver” node.
It used the wrong geometry, DOP geometry import (A) instead of (B). I thought that it is not meant to use this way and instead of fixing the path, I moved the (B) out of (A).

I also have a very strange behavior, that in some situations, re-simulate does not work. I tried to “reset simulation” manually on both solvers, but the replay didn't work. Only when I comletly switched off simulation caching checkbox.
Also forced re-cook by touching the nodes didn't work…
Only scene reload forced the correct re-simulation.
I will try to use disk cache instead of automatic memory cache.