Obi Official Forum

(07-03-2022, 01:13 PM)litefeel Wrote: [ -> ]Hi. did you find and fixed it.

Hi there,

So far I've been unable to reproduce any memory leaks, if you have a repro project please send it to support(at)virtualmethodstudio.com so that i can take a look.

I have run into the same issue I believe. It is affecting my ability to pass console soak-in tests for certification.

I have sent a repro project to the support email address.

If Obi solver is active with burst backend, native array memory gradually increases and never seems to decrease. I am using rope not cloth, but I thought it better to keep this post alive rather than create a new one.

[Edit: I have noticed also that there is no code compiled in the burst inspector for NativeArrayDisposeJob. I don't know if thisd is relevant but seems like it might be (see image)]

Hi Dominic,

Thanks for the repro project! Will take a look at it asap and get back to you.

(14-03-2022, 02:38 PM)josemendez Wrote: [ -> ]Hi Dominic,

Thanks for the repro project! Will take a look at it asap and get back to you.

Thanks for the quick response as always Jose.

Hi,

Leaving this here for future readers: As of Obi 6.4, yes there was a memory leak affecting particle collision constraints in the Burst backend. Thanks to SpiralCircus' report I was able to identify and fix it.

All you need to fix it in your project is to replace NativeMultilevelGrid.cs with the one attached.