I currently have a base spec M1 iPad Pro, and am thinking about upgrading to the new M4 iPad Pro sometime. I’m still relatively new to 3D modeling and sculpting (I’ve been at it for about a year or so), and so I want to make sure that my next iPad upgrade will give me plenty of headroom for larger projects. So my question is, would I be fine with sticking with a base spec for the new M4 iPad Pro with 8GB of RAM? Or would it be smarter for me to bite the bullet and pay more for a model with 16GB of RAM? The 1TB of storage would be useful, but definitely not necessary, I would be perfectly fine with 256GB of storage. Would the extra RAM make a significant difference using Nomad (for things like project size/quality), or would it be negligible or the same? I would greatly appreciate your advice on this! 
I use 2 Ipad Pros that are still without an M processor.
I definitely don’t need a new Ipad with M4 or 16 GB RAM for sculpting with Nomad Sculpt.
Who makes scenes with 30-40 million vertices and then doesn’t use the Quadremesher or decimation?
All the apps I know and use run smoothly and without problems.
I won’t need a new iPad until the apps also need the M processors, e.g. for ray tracing …
I thought about it for a long time, but I really don’t need a new iPad for sculpting, because I can’t even get close to the limits at the moment - even with the old iPad Pros.
Have you tried maxing out your M1 8gb (I’ve got the same model) I find it can handle 16million polys pretty easily with x, y and z symmetry all activated. With a large radius move brush it lags a bit, but is still totally usable. There is still over 4gb ram free.
With a 20 million object there is still 3.7gb free and performance is great. Experiment with your current iPad and that will give you a good idea of if you need an upgrade. (You might want to lower the quality settings (under post processing menu tab) to improve performance when using demanding models)
Thanks for the advice, I appreciate it! I did do a 3D model for a Sci-Fi fighter idea I had, and when I got to some point in modeling it in Nomad, Nomad kept crashing. I ended up figuring out that two of my little detail pieces were set at a voxel resolution of like 3,000 which was using up a bunch of RAM in the app apparently. When I voxel remeshed those parts down to 300, the RAM stopped redlining, and went down to like 3 or 4GBs in use. But that kind of got me questioning and wondering how much RAM some of my future projects would end up using up. Perhaps part of it also is that I’m not modeling efficiently, so I’m using more resources due to ignorance? This I’m assuming is definitely at least part of it, as obviously that situation I ran into was caused by out of control Voxel resolutions. If I can avoid upping the specs to a 16GB model, that would be preferable for me, because I’d be able to upgrade sooner. My current M1 iPad Pro setup is great for everything I’m doing, but I’m really excited about the OLED, hardware accelerated Ray Tracing, and the aluminum Magic Keyboard Case, so those are the main reasons I’m thinking of upgrading. I haven’t run into any performance walls yet, but since I’m newer to this, I want to make sure that’s not just because I haven’t started into larger projects yet.
It seems that an 8gb m4 iPad will probably only be able to handle scenes of a similar size to an 8gb M1, but it will have better performance. RAM mainly affects how many polys it can handle, the processor mainly affects how quickly you can do operations (sculpting with large radius brushes, remeshing, post processing, booleans etc).
Interesting… do you know what the overall scene poly count was when it crashed
Awesome. . How large of a scene can the 8GB M1 handle? I know you mentioned a polygon count, but newbie me is still trying to visualize what kind of a scene I could build with that. 
I still have a copy of it that isn’t fixed yet, so I can go and see.
I’m trying to open it, but it wants to crash every time I try to open it, so haven’t gotten a number yet.
You can get a feel for it by just adding various primitives, subdividing them until they have quite high poly counts. It could vary a lot: you might have one object, like a face with super high resolution and you could have 20million; or you could have 100s of lower resolution objects which all add up to many millions. I believe it behaves a little differently in these different sorts of situations, but I’m not quite sure how! Best to experiment by creating a big test scene and then trying various things like remesh etc. See what causes crashes and why.
Most of what I’m doing stays well under these limits so I have to deliberately try and crash it to find out!
It might have a bug issue that isn’t due to size.
Is the poly count that little number next to the object in the Scenes menu?
There is a number in top left of screen called scene vertices which is the total. Individual objects number is shown below
Sorry, yes, in scene menu it is next to object name
Ok, awesome! That’s what I thought that number was, but wasn’t entirely sure. . I got a screenshot of my stable version of the model I did. I can duplicate it and try remeshing parts at higher voxel resolutions. To test things out. .
Yeah, make sure you’ve saved a stable copy of the file and duplicate it before trying anything too extreme just in case!
I would say that total scene verts of 20million or more is quite high, so issues might start happening.
Looking at this sculpt you could definitely start bringing the vertices count way down. Have you paid for the recently introduced quadremesher tool? For sculpts like this it would be well worth it.
Once you’re happy with your hi res forms you can also use decimate tool to bring vert count way down (and this tool is already in Nomad, no extra cost)
I don’t think I did, I didn’t know there was a new tool you could buy. . I’ll have to look into that!
This will help with decimation
