Quote:
Originally Posted by F16_Petter
My suggestions for benchmarking track:
Contents of the track
1. Flying in clouds 1/2 or 3/4 (not overcast)
2. Flying over water
3. Playing around with both external camera and cockpit views
4. Shooting aerial targets, perhaps a bomber formation of 12-20 planes
5. Shooting / strafing ground target, truck or trains
6. Bombing something with secondary explosions, fuel tank or some large explosive item?
Atm my game is unplayable with my current rig so I cannot create a good track even with all settings on lowest... 
(lowest resolution and graphic options to lowest, anti epelectric filter off  )
I refuse to buy a new rig until I know what works best... and to do do that we need brave volunteers that show us what they get in a benchmark!
C´mon ppl lets make a track??
 |
It seems to me that the obvious way to capture all this stuff would be a formation of 109s attacking boats in a harbour. In about 5 minutes of track you could have:
- Medium level ingress above cloud
- Descent through cloud
- Combat over the coast (attacking targets in and around the harbour, facing AAA and then perhaps a flight of enemy fighters)
- Low level egress over both land and sea
The formation could split during combat so that the track includes sections with multiple aeroplanes in view and sections with perhaps just a pair flying home together.
The combat section of the track would obviously be very resource hungry.
However, some of the threads I've read suggest that track playback is considerably less demanding of system resources than actual flight (presumably because the simulation itself isn't running during playback), so it's not immediately obvious that this would be the best benchmark.
Perhaps we need some kind of repeatable* AI mission instead?
I wish Ubifail would hurry up and send me my (pre-ordered) collector's edition!
*In as much as the AI can be made repeatable - we'd probably need to apply statistics to get a useful benchmark if the AI exhibits random behaviour, which I expect it will do...