This CL rolls back the change from changelist 794576123 since the previous version was more robust and stable. Compared to the original algorithm, this changes handles degenerate and very large triangles as well.
BEGIN_PUBLIC
Restore mesh-SDF collisions.
END_PUBIC
PiperOrigin-RevId: 875716983
Change-Id: If0de41dc7bbd60460779821d2338cf2d661f40d5
Temporary changes to the octree:
- Changed frame from mesh to geom
Before change (tolerance 1e-3):
```
Simulation time : 2.34 s
Steps per second : 4275
Realtime factor : 8.55 x
Time per step : 233.9 µs
Newton iters / step : 2.26
Contacts / step : 3.69
Constraints / step : 14.76
Degrees of freedom : 12
Dynamic memory usage : 0.7% of 14M
```
After change (6 octree levels):
```
Simulation time : 2.24 s
Steps per second : 4472
Realtime factor : 8.94 x
Time per step : 223.6 µs
Newton iters / step : 2.47
Contacts / step : 3.37
Constraints / step : 13.49
Degrees of freedom : 12
Dynamic memory usage : 0.2% of 14M
```
PiperOrigin-RevId: 781019754
Change-Id: Ib15581244dfe9e571e1c6c2cac4a101cf3d8ba3d
Increased tolerance above zero in nutbolt.xml since now the two parts lock otherwise. Measured speedup of 2x in this example.
This solves the issue of scale-dependent step sizes. Tested on gears with 10cm diameter and 2cm thickness.
PiperOrigin-RevId: 584009449
Change-Id: Ied2f62dbbbc592470b91cf910f3553802a57c752
The following dependencies are disabled with a cmake patch file.
- Enoki
- OpenMP
- Assimp
- FCPW
PiperOrigin-RevId: 565352386
Change-Id: Icff9de68408e0afa2032bb83b24491fc4e41d181
chenlin
Alessio Quaglino