Page MenuHome

Box Trim Causes Crash
Closed, DuplicatePublic

Description

System Information
Operating system: Windows-10-10.0.18362-SP0 64 Bits
Graphics card: GeForce GTX 980/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.30

Blender Version
Broken: version: 2.91.0, branch: master, commit date: 2020-11-25 08:34, hash: rB0f45cab862b8
Worked: (newest version of Blender that worked as expected)

Short description of error
The Box/Lasso Trim tool in sculpt mode causes the program to crash. Task manager shows 99% CPU usage. This happens even if I restore factory settings.

Exact steps for others to reproduce the error

  • Start a new sculpting file (or add a mesh object and switch to Sculpt mode)
  • Select the Box Trim or Lasso Trim tool and use it on the object.

Related Objects

Event Timeline

Philip Holzmann (Foaly) changed the task status from Needs Triage to Needs Information from User.Nov 27 2020, 9:17 AM

I cannot confirm.
Does this only happen with high poly meshes? In that case you might be running out of memory.

Please run blender_debug_log.cmd from the Blender folder, reproduce the crash, and then upload the log files here.

It is kinda slow.
With the default sphere in the sculpting template, it takes multiple seconds here to box trim half of it.

Hi Philipp, thanks for looking into this. After giving it some time, the Box/Trim tool does it's job, but it does seem to take longer than it should. What concerns me is the fact that Task Manager shows Blender using up to 99% of the CPU while calculating. However, if I give it a minute it becomes resolved.
Thanks,
Mike

Philipp Oeser (lichtwerk) changed the task status from Needs Information from User to Needs Information from Developers.Nov 27 2020, 2:21 PM

Afaik, this is using exact booleans under the hood [which can be slow], so I assume this is more or less expected behavior @Pablo Dobarro (pablodp606) ?
Work on improving performance will continue, potential performance improvements are not handled as bug reports though.

Will leave up final decision to @Pablo Dobarro (pablodp606)

I have created report for this just now, haven't noticed this one. I think I have better steps to reproduce so I will merge this one.

I'd like to note that this bug was not about it actually crashing, but about it being very slow.
So this is not the same as the bug it was merged with.

However, I think that's just how it is, and boolean performance improvements will come in the future.

Ah I did not read whole discussion here. Yes in general performance issues are rarely bugs unless they are regression. And exact boolean is unfortunately slow.