18-02-2025, 09:03 AM
(This post was last modified: 18-02-2025, 09:06 AM by josemendez.)
Hi,
Just replied to your email.
For other users that may find this thread, the issue in this case was that colliders were not being disabled when no longer needed but instead were animated to have a size of 0. Unity does not support box colliders of size zero or negative, this lead to a large amount of warnings in the console and bad performance. Furthermore, contacts against these boxes were still being calculated and counting towards the limit of 266k contacts, even though none of them were actually progressing to collisions.
Also, 6 box colliders spanning the entire screen were used to keep the fluid inside a box, instead of a single inverted box collider. This also contributed towards generating a high amount of contacts.
Solving both problems greatly reduced the amount of contacts and allowed for more fluid particles to be used before missing collisions started to become a problem. Even then, this could be further optimized by batching together multiple blocks into larger colliders instead of having 1 collider per block.
kind regards,
Just replied to your email.
For other users that may find this thread, the issue in this case was that colliders were not being disabled when no longer needed but instead were animated to have a size of 0. Unity does not support box colliders of size zero or negative, this lead to a large amount of warnings in the console and bad performance. Furthermore, contacts against these boxes were still being calculated and counting towards the limit of 266k contacts, even though none of them were actually progressing to collisions.
Also, 6 box colliders spanning the entire screen were used to keep the fluid inside a box, instead of a single inverted box collider. This also contributed towards generating a high amount of contacts.
Solving both problems greatly reduced the amount of contacts and allowed for more fluid particles to be used before missing collisions started to become a problem. Even then, this could be further optimized by batching together multiple blocks into larger colliders instead of having 1 collider per block.
kind regards,