Hello

I have been working on a massive OpenWorld HDA for my team, which will assemble many other bricks, with organic logic behavior and rules driven by the artist.
Some of my multiparm have 3 levels of recursion with dozens of setting each.

- Have you been using method like Cacheif to only recompute the needed loop elements or other methods?
- I guess most of you will have choose PDG to handle the check that , straight out of the box right?
Will you convert the multiparm to a Dictionnary with Python, then to some global var to different HDA processors?

Hope it makes sense, happy new year!

vinyvince
- Have you been using method like Cacheif to only recompute the needed loop elements or other methods?

I don't think that's possible, CacheIf would not be able to hold a separate cache for each iteration
(the same way as you can't have independent DOP sim per for loop iteration)
at best you can get CacheIf to hold the cached geo across sequential iterations until invalidated
which would be sort of like a global cache per each Cache If node

vinyvince
- I guess most of you will have choose PDG to handle the check that , straight out of the box right?
Will you convert the multiparm to a Dictionnary with Python, then to some global var to different HDA processors?

you can use Attribute From Parameters SOP to help you with that, but then yes, you'd want to generate workitem per multiparm instance and add relevant workitem attributes based on multiparm instance parameter values

however I have not done TOP graphs with selective recooking requirement so I'm not sure what workflows are in place to ensure that workitems will not dirty downstream workitems if they result in exactly the same attribute values from your multiparm instances, but I assume should be possible