FSX, FPS & OPTIONS

[ppo 67]

FSX and some hard numbers. Part IIITest Setup:CPU: Intel C2D 8500 4.0ghzMB: Asus P5Q proVideo: XFX GTX285 1gigMem: Corsair DDR2 1066 (PL8)OS: Vista 64bit SP1FS: FSX acc packExtras:GEX, UTX, REX.(no AI used in the tests except Road tfc)1. FSX ACC pack and SP1 do not maintain a constant FPS even when nothing is moving: Situation description:B737 at end of rwy at KORD, NO AI, No tfc of any kind & no clouds. Every 30 seconds or so, the FPS goes into a dive during another 30 seconds.The dip is significant but will depend on many variables: ur PC and the FSX environment.The regular 30 second dip applies to SP1 and SP2. I have tested both.It also applies to XP SP3 and Vista SP1 64bit.I have lowered the FSX visuals to minimum and checked off everything but it still dips.After 60 minutes of flying, it still dips.That is very important to realize if you do any benching or just flying. I don't have the knowledge to explain this; FSX is obviously doing some local processes which affects the FPS readings. Here's a picture to demonstrate. (5min sample)2. The following has been done to evaluate the impact of the settings on FPS. Your mileage may vary.Reference flight:Default 737 end of rwy KORD. Outside view with Airport in the background.Resolution is 19X12X32 and nHancer settings: AA4 AF16. Video driver 182.06whqlRef. settings: Water effects: Low2X Filter: Bilinear Advanced animation+flare: on Cast shadows on a/c + lights on ground: on Most settings at MAX unless indicated.I.Autogen & Scenery complexity a. Autogen & Scenery complexity at Extremely dense : 46 fpsb. Autogen & Scenery complexity at V. dense: 56 fpsc. Autogen & Scenery complexity at Dense: 72 fps d. Autogen & Scenery complexity at normal : 82 fpsII. Ultimate TerrainNo lights in UTX with Autogen/Scenery at Dense: A1. NO LIGHTS (road-intersections): 72 fps A2. With UTX lights : 44 fpsIII.Water DetailsWith Autogen/Scenery at NORMAL: 82 fps that is with Water at Low2X.If Water details is set at NONE = 102 fps. And from Water Low1X to High1X, there is no change in FPS. If it is set at LOW2X, we get 82, if we set it at MID2X, we get 82 also. If we set it at HIGH2X, we get 75 and for MAX2X we get 60.IV.Level of DetailMesh does not affect FPS but Levl of Detail does: LOD Max = 82; LOD Med = 90; LOD Small = 96V. Road trafficSame baseline as the ref. flight mentioned above with NO road tfc, we get 82. If we bring it up to 10, FPS reduces to 63, to 20 it reduces to 58 and 100% road tfdc we get 48 fps.VI. Ship trafficNo effect even at 100%VII. Pleasure craft trafficNo effectVIII. Airport densityAgain same ref. flight and FPS at 82. Is AP density set at MIn, we get 59 fps, at LOW, we get 46, at MED, we get39, at HIGH, we get 36 and at V HIGH we get 31 fps.IX. Ground ShadowsChecked and no effects!!! will test further.X. BloomChecked and FPS goes from 82 to 46.Some observations:1. if you do any benching on FSX, make sure you consider the 30 secs dips every 30 seconds;2. the biggest FPS cruncher: BLOOM -36FPS in second place: UTX lights (All) with -28 FPS in third place: AP density (at MIN: -23FPS) Water Details is also hard on FPS = - 20 FPS.Conclusion: I'm not sure there is any conclusion here. Obviously, your results will vary. What we need to remember is the 30 second cycle and that MS supplied us with many settings that need to be adjusted to your particular hardware and I mean not only for eye candy here: know the limits of your hardware.Hopefully that may prove to be helpful to some.Pierre

[Bert Pieke 7,584]

FSX and some hard numbers. Part IIITest Setup:CPU: Intel C2D 8500 4.0ghzMB: Asus P5Q proVideo: XFX GTX285 1gigMem: Corsair DDR2 1066 (PL8)OS: Vista 64bit SP1FS: FSX acc packExtras:GEX, UTX, REX.(no AI used in the tests except Road tfc)1. FSX ACC pack and SP1 do not maintain a constant FPS even when nothing is moving: Situation description:B737 at end of rwy at KORD, NO AI, No tfc of any kind & no clouds. Every 30 seconds or so, the FPS goes into a dive during another 30 seconds.The dip is significant but will depend on many variables: ur PC and the FSX environment.The regular 30 second dip applies to SP1 and SP2. I have tested both.It also applies to XP SP3 and Vista SP1 64bit.I have lowered the FSX visuals to minimum and checked off everything but it still dips.After 60 minutes of flying, it still dips.That is very important to realize if you do any benching or just flying. I don't have the knowledge to explain this; FSX is obviously doing some local processes which affects the FPS readings. Pierre

I believe the ACES team explained that every minute FSX redoes the lighting of the scene.Maybe that is your 30 second event. In which case, it is by design.

[ppo 67]

Tks for the info Bert. In that case, they're off by 30 seconds

[Guest JimBrown]

Not sure I can agree with that...If you look at your chart, the dips start at approximately 60, 120, 180 and 240 seconds. That corresponds with the "lighting refresh every minute." It appears that the refresh takes about 25 seconds, although there could be other things being done as well.Is the dip seen when the target framerate is set to something?...jim

[ppo 67]

..JimWhen you start FSX the first wave lasts 57 secs but afterwards it cycles every 30 seconds (plus or minus 1). (60+30=90+30=120 etc...).The dip can be made to disappear yes if you lock the frame rate at say 30fps and your PC is able to generate more than 30fps. But if not then the cycling is seen.

[Mace 1,902]

To add to the mix, I have never observed regular frame dips like that in any testing I have ever done.Are you absolutely 100% sure that you didn't have any background processes going then? Did you check with an app like Process Explorer, to see what was going on then? What about OS file indexing?

[ppo 67]

No indexing in either XP or Vista. Running the bare essentials: 30 processes for XP and 40 for Vista. Will check with Process Explorer

[Bert Pieke 7,584]

Tks for the info Bert. In that case, they're off by 30 seconds

Your chart shows a period of about 60 sec :(