Crash towards end of wrapping

had a mesh of 1 million faces on a scan mesh, needs all the detail but not the bumpyness of the scan mesh its for extreme closeup

was going ok but when it gets towards the end of wrapping after a couple of hours it crashes tried it twice now :frowning:

Hello,

Wrapping is very memory consuming, since it uses optimization approach. Most likely the problem arises due to a lack of memory. I’m affraid that 1000K basemesh is too heavy for Wrap. It only works nicely for meshes under 100K polygons.

If your basemesh is created using a subsequence of mesh subdivisions, I would recommend to wrap a lower subdivision level (under 100K) then return back to the highest subdivision level (for example ZBrush allows that) and use ProjectGeom node or projection in say ZBrush to project small details. In practice there will be almost no difference in result if you did it by wrapping but it would save a lot of time.

Hi!

I was running into the same issue, freezing around 34Gb memory usage. I have 128Gb though, is there a way for wrap to access all the memory I have, or should I try to trim the high res mesh further? It would be great if we could use every bit of memory before crashing ( especially in the years coming, I suspect more and more complex stuff will be made by peeps ).

cheers,
b

Hi! :slight_smile:

And sorry about the delay, I’ve solved the issue for the time being. The problem (as you mentioned) was that both my scans and basemesh was pretty dense. Lowering polycount solved the issue, but required some post-wrap craftsmanship. The scans weren’t perfect anyway, so the manual labour was expected. The original basemesh had roughly 400k, but using lower subd levels was ok.

Hello kissb!
How many vertices in your basemesh (not scan)? As Andrew already mentioned Wrap is not tested on basemeshes which contain more than 100k vertices. Can you send basemsh and scan to us? (If you can i will send my email via PM) We can sign NDA if it needed.
Sometimes last stage of wrapping consumes several times more memory than previous stage (4-5 times).

It would be great if we could use every bit of memory before crashing

You can try to adjust sampling final parameter http://www.russian3dscanner.com/docs/Wrap3/Nodes/Wrapping/Wrapping.html#parameters-tab. Bigger values should lead to less memory consumption. But it can be quite painful because each try will require a lot of time.