Is it normal for Blender Scuplting to lag at 1-2 million tris?

Hey guys,

As most of you i got my self a new rig with 5800x ryzen 7 + 3070 card. Now I see that it lags out as soon as a draw a stroke on 1+ mil tris, I see heard that zbrush can handle lot better but I dont have it as of now.

I wanna know if you guys are sculpting skin details/ skin pores/ wrinkles etc, do u ever use blender for that? if so how much more resources does it really need?

Thanks!

Thats not normal. Whats your setup? Any modifer? Dyntopo or something? With 10mio its still sculpting great and even more…

My computer lags at 500k Using an old Gen2 Xeon. Dynotop.

Oh! I had subD on dyntopo’ed object! :grimacing: :man_facepalming: Thanks!

Dyntopo performance is currently not good but its been worked on and in 2.93 solved i hope. In the meantime its advised to use the voxel remesher that will definitely not lag at 500k.

That sounds wonderful. I depend on dynotop, it’s my main sculpting tool. If they have fixed the performance it will be even better. I was considering 3dcoat’s sculpting tool. But I have to make an account just to use the trial version. So I am not going to get into testing it.

1 Like

Its in steam if u use that

I can’t use steam because of the hackers. And now my account is permanently locked.

I found out that dyntopo is depreciated. Performance problems. The way to sculpt is using voxel remesher.

Do you have a link to where it’s confirmed that dyntopo is officially deprecated? I see it mentioned as a possibility in the dyntopo roadmap but haven’t turned up anything saying it’s confirmed…?

Dyntopo has been under development by the original author recently.
(https://devtalk.blender.org/t/2021-01-19-tuesday-talks-module-meetings/17132)
That started not long ago and afaik that has been much later then the roadmap document has been written.
Once review and rework have been finished it should be in much better shape.

I think that makes it less likely that we will see this feature be deprecated.

The patch is finished and under review. Expect it for 2.93 or later. Dyntopo is strong as ever after that.

Yes but the patch will have to be split and its already mentioned that some parts will still need adjustments. But yeah I’m also looking forward to it being rolled out. :slightly_smiling_face:

I didn’t want to say it was officially depreciated. But that is the plan. I don’t think they are going to keep it. But maybe, it has lots of users. In the mean time, I think the voxel is worth looking at. I’ll be doing work in voxels for now. But maybe they can find a way to fix the performance of dynotop? They said voxels is better at sculpting than dyntopo. People reject it I think, because they are used to dyntotopo, rather than voxels. 3d coat uses voxels.

Voxels is not new in blender, it has been around from 2009 I found. Universities have been using dyntopo in the same way voxel is used, they sculpt sith dyntopo, then turn it off and sculpt. I think dyntopo is better, when used this way. It gives similar performance as voxels when used that way. I’ve tried it. Except you get dyntopo topology and workflow. Problem on my machine, is slower switch dyntopo on and off.

I want them to keep dyntopo.

The current sculpting in blender heavily based on pixars openvdb, its initial openvdb release is from 2012 and its official implementation/embedding in blender is not even 2 years old .The voxelreseshing algorithm is also part of openvdb. Dyntopo has been there in blender much longer, and its continous remeshing makes it shine in several areas, but its usage has grown and some initial code decision turned out to be a misfit because of that.

With the new patch, I really think it will stay, it will just be adapted over time here and there to make the two approaches complement each other better.

I hope they don’t deprecate it before introducing a better solution because right now voxel remeshing an entire mesh just so I can have detail on one small portion of it makes my computer very sad.

However, it’s entirely possible that I don’t understand the intended workflow and/or how to optimize for using voxel remeshing.

I think dyntopo makes better form. The form is better than using voxels. This makes dyntopo better at sculpting.