Having already blogged about transparency issues in Acceleration/SP2, here is a list of some additional issues we have identified. Note these are in no particular order and with no attempt to be exhaustive, this is just a list of what I know and can address authoritatively at this time.
1) River Thames. Longleat is in the same QMID as the Thames, and a re-export of that QMID on 8/22 picked up the pre-SP1 data. It appears only 2 slope problems exist, as compared to how many slope issues were in RTM at this location, but this is an unfortunate regression it is true.
2) SmallPartRejectRadius tweak. It appears that batched objects are not respecting this setting. It’s not clear this is causing any performance issues since the objects are batched, but it is still a change from SP1.
3) VC Lighting in daytime. There was a change to the aircraft.cfg for the following aircraft:
o Maule with skis
These aircraft can be fixed by uncommenting the line containing the fx_vclights keyword. No code change was made as part of this change, so that should revert the aircraft to SP1 behavior.