I moved from an A15 iPad Pro to the newest M4 one just for Nomad (because damn it’s fun and addicting), but unfortunately I have not seen that much performance improvements.
On a 5.5 million poly object, the framerate is always around 20-30 when rotating it, doing something quickly occasionally drops it to 5, and during sculpting it feels as it’s 5-15fps even though the debug framerate says it’s a bit higher.
I haven’t enabled any extra graphics options, just one of the built-in matcaps. I do run with dynamic tesselation always on because the brushes behave better using it. I recently switched from screen based dynamic tesselation to constant but it did not help. Using constant however, I actually got the sculpt down from 5.5 to 3.5 million polygons, but again the feeling of brush performance was only marginally improved.
I’ve tried running with dithering off and smooth shading off, but that did not help. Can I do anything else?
Now, my bet is that I will be told to reset settings to default, but if that doesn’t help I have made a lot of small changes here and there, so I’m wondering if there’s a way to export and import settings so I don’t have to start all over again?
Attaching a screenshot where it’s quite apparent how poor the brush performance was at 5.5 million… those are supposed to be continuous strokes, not dots…
What’s the resolution in postprocess menu, it should be no higher than 1.25.
Dyntopo should be set to quality to avoid the dot thing.
Do not display wireframe when sculpting with dyntopo.
Brush parameter stroke spacing shouldn’t be too small (in case you changed it).
Either doesn’t impact performance, and smooth shading impact is minimal.
The values shown in stat is not the polycount but points count, on a triangle meshes the ratio is usually twice triangles than points. On a quad mesh it’s usually around the same.
Interesting post.
I also still work with an older 12.9 inch Ipad Pro with A processor.
I work at most in a polycount range up to 5-6 million - but only for a short time, otherwise I always decimate / quadremeshe each project below 1 million vertices.
Apparently a new Ipad Pro with M4 doesn’t bring any significant gain if you use Nomad.
Then I’ll wait until it breaks before buying a new one .
Ah, are those the small specks you only see in Blender?
I was only doing it for diagnostic purposes.
Just tried to reset it and the number did not change so that means I must have been using the default value.
Thanks for the rest of the information!
Interesting. That’s when my old iPad choked and I decided to get a new one, as I noticed that Nomad’s limit was 10 million.
But well, I’m getting a sense of deja vu now because:
Today, working on the same model, I get pretty steady 60-90fps, including during brush strokes and I can even turn on the curvature post process without issue.
However, weirdly I notice that Nomad has less than 7gb of ram available today.
So, that makes me wonder… was it a background process that was running for a few days slowing Nomad down? Possible to bloody AI thing that was indexing my photos perhaps, even though I tried to turn all that crap off?
I know I’m just repeating what everyone else is saying here, but this all sounds like a dynamic topology thing. If your brush is set too large, you’re basically telling your iPad to add or remove potentially hundreds of thousands, or even millions of vertices at once. This is one scenario where you’re definitely going to be CPU/GPU bound before your memory becomes a bottleneck.