Hi!
we are noticing that certain VRmeshes (for example ones that have a very few triangles per voxel, thus high number of voxels) will generate segmentation faults when rendering Linux.
I assume that the high number of voxels is making VRay needing to allocate too many regions of memory and this eventually will cause VRay to crash...
This same vrscene renders without a problem in Windows.
so that brings up a couple of questions...
Is there a known limit to memory regions allocated in Linux? And how many regions does a vrmesh try to allocate?
Is the number of regions of memory allocation dependant on number of voxels? Number of triangles? Number of triangles per voxel? Preview faces? A combination of all of the above?
I know that it is not recommended to have a low number of triangles per voxel (as this would generate a high count of voxels), so is there any way to run analytics on an existing vrmesh? See how many triangles per voxel, or voxels per file, are there? A python way to run that would be optimal
we are noticing that certain VRmeshes (for example ones that have a very few triangles per voxel, thus high number of voxels) will generate segmentation faults when rendering Linux.
I assume that the high number of voxels is making VRay needing to allocate too many regions of memory and this eventually will cause VRay to crash...
This same vrscene renders without a problem in Windows.
so that brings up a couple of questions...
Is there a known limit to memory regions allocated in Linux? And how many regions does a vrmesh try to allocate?
Is the number of regions of memory allocation dependant on number of voxels? Number of triangles? Number of triangles per voxel? Preview faces? A combination of all of the above?
I know that it is not recommended to have a low number of triangles per voxel (as this would generate a high count of voxels), so is there any way to run analytics on an existing vrmesh? See how many triangles per voxel, or voxels per file, are there? A python way to run that would be optimal
Comment