[Bug] Applying Pin Configuration step have become really slow between v.1.0.13 and v.1.0.14

[PhysicalTesting]

Summary

I tried to test the solver for one of my characters to simulate the muscle/adipose tissue, which require a shell like structure and a lot of pinning. It works, but the pinning have become slow for a character consisting of around 25.000 vertices.

The solver also thinks it's slow, as it shows a timeout message during 'applies pin configuration'. At the start I thought the pinning process had crashed or just stopped responding, but the console kept updating and the process showing '72% Applying pin configuration...' until it finished and able to simulate.

I thought my model had some issues since pinning was so slow. I tried to track down the issue and managed to recreate the 'time out' consistently in a new file by having a really detailed mesh, just a ball with a core, totalling around 100000 faces with a pinned core, actually just an advanced model so it generates a lot of vertices, triangles and tetrahedras.

That lead me to do a comparison between 'Blender Add-on 1.0.13 (addon-2026-07-01-0327)' of the solver and the latest version 'Blender Add-on 1.0.14 (addon-2026-07-14-0335)' with the pinning, and the newest version is significantly slower compared to the older version 1.0.13 that did the pin configuration almost instantly.

From the test, it seems the pinning took a big hit in performance between the two versions. 🤔

Steps to Reproduce

1. Have a advanced manifold mesh set to solid so the solver generates a lot of triangles, vertices and tetrahedras.
2. Add a pin group so parts of the mesh is pinned.
3. Animate the mesh and re-capture all deformations
4. Click transfer, it goes really slow, it reaches 'Applying pin configuration', the Timeout eventually shows, but it should mostly finish.

Expected Behavior

Keep its speed when applying pin configuration as it did in version 'Blender Add-on 1.0.13 (addon-2026-07-01-0327)'

Actual Behavior

Applying pin configuration have become really slow in version 'Blender Add-on 1.0.14 (addon-2026-07-14-0335)', often shows Timeout even though it eventually finishes.

Environment

Windows 11
Blender version: 5.2
Solver version: Blender Add-on 1.0.14 (addon-2026-07-14-0335)

Contribution Agreement

• I confirm that any code, patches, scripts, attachments, or other copyrightable material I include in this issue may be used, reproduced, modified, and distributed by this project under the Apache License 2.0 or as otherwise described in CONTRIBUTING.md.
• This is a bug report from the community (please provide a reference below).

Community Reference

No response

[ryichando]

Sure! I will take a look at it. May I ask you to post a reference .blend file so that I can quickly reproduce the issue and fast start working on it? You can zip it first then drag-drop in the reply.

[PhysicalTesting]

Yes of course, I got the test file right here. ^^

ZOZO pin timeout test.zip

[PhysicalTesting]

Does this addon use Blenders deform modifiers for binding the tetrahedralized mesh to the graphical mesh?
I know that Blenders deform modifier gets exponentially slower when binding precision is increased for a mesh.

[ryichando]

No. The addon does not use Blender's deform modifiers for binding the tet mesh to the graphical mesh, so the slowdown behavior you pointed out about doesn't apply here.

[ryichando]

Hi! I've made big architectural changes to the way pins are handled before this issue is posted, not just on the engineering side but also in the underlying mathematics. With the latest version on my end, I was not able to reproduce the issue. It may have already been silently fixed, as shown below:

So I decided to tentatively mark this as worked. The new version will be delivered in the next release. After the release, and if you still encounter the issue please get back to me here.

[PhysicalTesting]

WOW! Thank you so much for looking into it, I really look forward to try it out! 😄

[ryichando]

Delivered in this release

https://github.com/st-tech/ppf-contact-solver/releases/tag/addon-2026-07-27-0038
https://github.com/st-tech/ppf-contact-solver/releases/tag/2026-07-26-22-53

[PhysicalTesting]

I tested the animated solid soft ball and the timeout does not show any more.

I tried with a new shape with crevices, something that can simulate soft characters or just any cloth or softbody with folds. I made a cylinder with deep grooves and a internal core, where the internal core works as bones, as if it's for a character. The model is rigged with the internal core pinned, along with the outer ends as well as the deep grooves. The rig got a simple animation where each bone bends slightly while it turns over. It uses blenders own cloth solver (from frame -60 to frame 5, cloth sim disables at frame 3) as a mesh preparation method, to keep the mesh inflated while preventing the mesh from clipping through itself.

It does take a really long time to apply pin configuration, especially when finishing pin configuration where it generally takes around 9 minutes before it finishes. It also shows a 'Operation time out' message like before. Only once, it showed a timeout during tetrahedralization. But in all instances it does finish the pin configuration.

Maybe this shape was just demanding, I tried version 1.0.13 (addon-2026-07-01-0327), it was a bit slower to prepare it compared to the ball, but the old version was very fast.

This might need its own thread:

But after frame 26 the simulation fails with different intersection detected messages.

I can't really see what should intersect when inspecting the mesh from the inside. Just to be sure that the pinned crevices didn't cause the intersection errors when the rig bends, I removed the pins in those crevices, but the simulation still fails.

Removing the pinning cause no intersection errors and the simulation runs fine.

The blend file for testing, blenders own cloth cache files should be with the blend file. I have uploaded it to mega:
https://mega.nz/file/lntyGTJI#QpBL2ZwuigWqtnYXhGgPLdOiNSkEPrwqMh5gqAnmvhE