OK Guys,
Bit of a strange one.
Can’t really discuss too much detail as this this relates to the ‘secret sauce’ in my code.
But this is how I can frame it.
I have been using 3x GH Record components holding txt and curve data up to 500 steps deep in a setup with 2x Anemone nested loops (yep it is as weird as it sounds)
What I have noticed since this code was written was a memory leak - when this looping is running the memory usage of Rhino just ticks up, 3GB to 10+GB over the course of 30 min or so.
Stopping and restarting the code (which resets the record components to empty) doesn’t clean up the memory usage and it will tick up from where it left off.
As you might expect this code will slow to a crawl after a period of time.
Until now I’ve not looked into it very much as the code was working.
Well this week I’ve looked into it and what might be leaking memory like a sieve.
First port of call was the GH Record components, which I’ve replaced with some “roll my own” GH Python code using a sticky_dict. This wasn’t the easiest as I need to save “groups” of curves as part of the process.
Well, I got this to work on R7 and the memory leak seems to have been solved. The memory usage tops out at about 3GB and will drop back as the process stores less data/will drop right off if I clear these GH Python sticky_dict.
Now all running GREAT on R7 and has improved the overall performance of my code a LOT BUT load it into R8 and as soon as I run this new code the Rhino/GH/Custom UI locks up. I can see my code running fine in the Rhino viewport but I can’t get control back.
The only way to kill it is from Task Manager.
I will try to produce a ‘something’ that shows the issue. But before then any ideas brains trust?
Cheers
DK