CloudCompare forum

I have a huge pointcloud, almost 2 billion, which fills about 70GBs of memory, I have 115GB RAM. The cloud load fine, of course the performance is pretty slow, but that is expected, the problem is I need to segment this cloud into multiple smaller ones, with each segmentation step using the segment tool, the RAM consumption almost doubles in size, untill it crashes. I am running this on Ubuntu and using CloudCompare v 2.11.3. My question is why this happens, and is it possible to stop it?

By 'segmentation step', do you mean you are closing the tool each time and re-opening it? Because in this case, at the moment the cloud is split into 2 parts, there's a copy of some of the points to a new cloud (before removing them from the original cloud). So there's a temporary memory increase. You also have the option to permanently deleted the invisible points (with the small bin icon). That might help reducing the memory load.