MultiJoints issues

[tsn]

I'm currently working on some tests for a more complex rbd simulation with extensive use of MultiJoints. Unfortunately, the new new version (released on July 14) is a huge step backwards:

- Simulation times are 1.5x - 8x longer compared to RF5.0.0.
- With simple scenes, CPU load is approx. 50%, with more complex scenes it'S just around 12%. In RF5.0.0 I had constant loads of 80-90%.
- Estimating forces is still nothing more than a guess and there's absolutely no indicator for finding a working initial value.
- The forces, displayed in the statistics panel, have obviously no relation to the setting under @ Max force, because there the used forces are around 1 - 10% of the values from "Statistics".
- There seems to be a new way how forces are calculated, because the scenes from the previous version now yield completely different results.
- Some joined objects show heavy intersections (but it's not possible to find out why).
- With Quality = High, simulations aren't possible at all, because simulations times are incredibly long.

In the current state, MultiJoints are actually not usable and that's very disappointing. I hope that there will be some major improvements soon.

Best,
Thomas

Version 5.0.1 on OS X 10.6

[agus]

Thanks a lot for your feedback. We will work hard on it.

[nuria]

RF 5.0.3 is now out and it has substantial improvements in all these issues.
Just the one concerning the forces displayed at the statistics panel, remains. Unfortunately, though internal multi-joint force values are correct and respond appropriately to user's input parameters, they are not showed properly. This will be fixed for next version.