Hi
First of all , again many thx to fix the compatibility issue we had between Krakatoa 1.5 and new builds of TP4
We have realized that the automatic seed increment does not work when caching partitions from thinking particle 4.For now I think we can use the Krakatoa_SeedIncrement_UserFunction() but would be cool if you guys could fix it in the core.
Krakatoa version : V1.5.2.39428
Best regards
I just tried partitioning with Box#3, and the seeds did not randomize there either. Perhaps something odd about partitioning in general?
Edit: Yeah, just tried both local and deadline, neither method incremented the seeds.
Ok, testing…
For Box#3, I was adding a Random Vector Spherical Volume to the Position and outputting Position.
All my partitions ended up co-incident.
Manually changing the seed and saving the PRT got the results I was expecting.
What version of Box #3 are you running?
Looks like Oleg broke the MAXScript code in 1.5.
The function .getName() is not defined anymore (it was used to grab the name of the operator), and the following code misbehaved too:
a = $Data_Operator_01
–> $Data_Operator:Data Operator 01 @ [0.000000,0.000000,0.000000]
classof a.SubOperators[2]
–>randomReferenceTarget
classof a.SubOperators[2] == randomReferenceTarget
–>false
(!!!)
So I had to switch to the following code:
if matchPattern ((classof subOp) as string) pattern:“randomReferenceTarget” and IncrementBox3Random do
(
local theSubOpName = try(subOp.getName())catch(subOp.name)
…
First, it checks the classes as strings and that works.
Then it tries to get the name using the old getName() function and if it fails, it tries the new .name property.
This works now in 1.5.2 with Box #3 1.5, but I also hope it would work with the older Box #3 builds that behaved differently…
Will post a fix whenever I have it (have to modify the Deadline code, too).
As for Thinking Particles, we know that MatterWaves is not supported (since it has no MAXScript access to the Seed).
I will try to take over Darcy’s machine for half an hour to see what exactly is going on since I have only TP 3 here.
I tested a simple TP3 scene with a PositionBorn op and got correct random seed incrementation.
I was unable to get whatever TP4 version we have to run (not even on Darcy’s machine), so I am unable to test it until TP4 is released and we purchase some.
I could send you a simple script to check what properties Krakatoa is trying to read might have changed in TP4 on your installation… Other than that, I cannot do much at this time.
Ok, we managed to get the TP3 scene I created into a copy of Max with TP4 which, while not really licensed to edit the TP system, could at least run the pre-created file.
With a simple PositionBorn creating default particles, it partitioned correctly and created 10 times more particles with random positions.
As I mentioned, MatterWaves is not supported in any version due to lack of MXS exposure.
Other operators might have been affected by changes in TP4, but we have not tested yet.
Can you test a simple PositionBorn and tell me whether it is broken on your system or not?
Also, can you look in the Krakatoa Log after running partitioning of your more complex scenes and copy&paste the reports about what is being affected by the Seed options?
Thank you!
I know this is an old post, but wanted to mention that with Krakatoa_SeedIncrement_UserFunction() it’s possible to look for an integer node by name, and modify the integer as a seed… which can be plugged into a matterwaves’ random input, effectively ‘fixing’ the issue. I’ll post up a script after TP4 comes out (which should be soon).