The charts in Tom's Hardware Guide suggest that there is no appreciable difference between an 8800 GTS and an 8800 Ultra in FSX. This seemed strange, even given FSX's CPU-dependency, and I have always wondered whether these tests used SP1 or not. Curiosity finally got the better of me and I have upgraded my 8800 GTS 320 Mb (factory overclocked to 580MHz) to an 8800 Ultra (factory overclocked to 655 Mhz).It is notoriously difficult to convey any idea of relative smoothness in FSX because framerates are seldom constant, except at high altitude or in areas of low graphics complexity, where they don't matter. But to give some idea of what I mean (though without any pretence at being scientific about it), I have conducted two comparisons using identical settings in pretty demanding situations:1. Level-D 767 parked at EGLL Gate 4148800 GTS: about 18 fps8800 Ultra: about 26 fps2. Level-D 767 on finals to PHNL8800 GTS: about 17 fps8800 Ultra: about 25 fps.(Settings used: driver version - 163.71 beta; Nvidia control panel - all "default"/"let application decide"; FSX display options: target frame rate - unlimited; resolution - 1920x1080; filtering - anisotropic; anti-aliasing - on; global texture resolution - very high; lens flare - on; advanced animations - on; light bloom - off; aircraft casts shadow on the ground - on; aircraft landing-lights illuminate ground - on; aircraft casts shadow on itself - off; level of detail radius - large; mesh complexity - 100; mesh resolution - 1m; texture resolution - 60cm; water effects - low 2.x; scenery complexity - extremely dense; autogen density - normal; ground scenery shadows - on; special effects detail - medium; cloud draw distance - unlimited; detailed clouds - on; airline traffic density - 0; general aviation traffic density - 0; airport vehicle density - low; road vehicles - 0; ships and ferries - 18%; leisure boats - 18%).(Computer: Dell Precision 690, 2 x 5160 Xeons, 4GB RAM, 15k RPM SCSI drives).(Operating system: Vista 64 bit "ultimate").I also conducted the same comparisons but with EVERYTHING set to "max" (light bloom - the whole lot). In both examples, the 8800 GTS gave about 4.8 fps; the Ultra managed 6.5 fps: an improvement of about 34%, but still completely useless.Incidentally, I also tried changing the resolution but this made no real difference to the relative performance. So I do not think the difference is attributable to the extra RAM in the Ultra.As a counter-point to those numbers, my impression from general use of FSX is that the Ultra typically produces about a 20-30% improvement in overall framerates.I do not for a moment suggest that the 8800 Ultra represents good value for money: obviously, at

I don't thin Tom's would have been using the LDS767 for a comparison!Did you take those numbers from the VC?It would make sense if you did as the VC in the 767 is very taxing in FSX.In this situation I would totally expect what you are seeing as the Ultra can process the added stress from the LDS767 VC / more detailed model better than the GTS can given it's higher performance.Try repeating your experiment with a default aircraft. Then I think you will be more in line with what Tom's did.Your test is a good one though as most of us here will want to end up using a complex addon in FSX anyway. In this case it may help someone decide which of the GTS/GTX/Ultra cards they would like to get.Good stuff......Glenn

The second scenario where you had everything to the max was more interesting. Unless there is an "adjustment" that will help, it shows that the performance ceiling remains even on a quite expensive system. Perhaps those who have high-end expansive systems could repeat the simple test scenario to see if someone could reach say 20fps with everything set to the maximum. There may be a CFG setting that will now benefit those with resources beyond a certain level whereas that setting seemed to offer no advantage before.Regards,Dick BoleyA PC, an LCD, speakers, CH yoke

You're right: I was using the VC cockpit. I now wish I'd tried a default aeroplane while I still had the 8800 GTS in the machine. My gut feeling is that this cannot account for the difference, or at least for all of it. But if I can face the hassle, I might re-install just to find out. Tim

I just tried this test with my 8800 GTS and I managed a whopping 8.9 FPS. ALthough I read your post wrong and parked at gate 114. Let me see you get a 737 out of that parking spot ! LOLThe only difference in our settings is that I have traffic maxed out and Ultimate Terrain and XGraphics takes a CPU hit too but I am a GA flyer and rarely go to these big airports. Now I am curious as to what it will take to get to 20 FPS there. I swear I spend more time messing with stuff than I do flying. Ok.. I turned off ALL aircraft traffic (road, boat etc at 60%) and I got 31FPS but that airport was like a ghost town. I feel a little better now though but I wonder how you do with traffic maxed. My 8.9 FPS I had all traffic at 100%. I wanted to go the Ultra route but couldn't afford it (read ..wife would have killed me) ;)EDIT: After I sat there for a few minutes, the FPS sank down to about 18. It must have taken awhile for all of the textures to load.

I love such controlled experiment focusing on FSim.Thanks Tim for taking the time to do that. I appreciate it very much.Manny

Tim,I just ran your EGLL Gate 414 and settings scenarios on both mine and my son's computers. Our computers are basically the same (C2D Allendale, 2GB RAM, Gigabyte DS3 mobo, XP Pro) except he has my "old" 7900 GTO 512MB and I have an 8800 GTS 640MB. While my CPU normally runs at 3.2GHz+, I set it to 3.0GHz the same as my sons for the test and we also used 1680x1050 as the resolution, as this is the max my LCD panel can run at.Results were as follows:1. Level-D 767 parked at EGLL Gate 414, your specified settings: 7900 GTO - 24 FPS8800 GTS - 25 FPS2. Level-D 767 parked at EGLL Gate 414, EVERYTHING set to "max", including light bloom:7900 GTO - 5.6 FPS8800 GTS - 7.2 FPSWhilst the performance crown went to the 8800 at the everything max (albeit totally unplayable) setting by about 30%, that advantage dropped to a measly 4% at playable settings. If I then factor CPU-intensive AI into the equation, as I usually run 100% airline AI with Ultimate Traffic-X, that margin drops even lower again as the GPU becomes even less of a factor in resultant FPS.In my case at least, my base performance with a 8800 GTS is quite comparable to yours with your 8800 GTX Ultra, so I don't think a faster graphics card would make things any better for me. Perhaps it's an XP versus Vista thing??Gary

It's funny how FSX's performance is affected by seemingly trivial differences. I expect you've got faster RAM than me and as you point out, XP seems to give better performance than Vista. What your experiment shows is that I need an Ultra to get the sort of performance your machine gives you with an 8800 GTS. What I don't think it shows, is that you wouldn't get the same proportionate improvement as me in graphically-intense situations if you used an Ultra instead. If my experience is typical, you'd be looking at about 34fps at EGLL gate 414.Tim

Or that if I "upgrade" to Vista, my FPS with my 8800 GTS may too drop down to what you were getting with your GTS. When Vic (vgbaron here) did some test runs for me of XP versus Vista, we noted about a 10% FPS penalty for running Vista. Gotta love the quirks of computers and software :-lol.Gary

Was that the default EGLL Tim or the Aerosoft Heathrow scenery?

>Vic (vgbaron here) did some test runs for me of XP versus>Vista, we noted about a 10% FPS penalty for running Vista. I sure hope that changes with Vista SP1. RhettAMD 3700+ (@2585 mhz), eVGA 7800GT 256 (Guru3D 93.71), ASUS A8N-E, PC Power 510 SLI, 2gb Corsair XMS 3-3-3-8 (1T), WD 150 gig 10000rpm Raptor, WD 250gig 7200rpm SATA2, Seagate 120gb 5400 rpm external HD, CoolerMaster Praetorian

Default scenery. I can't get Aerosoft Heathrow 2008 to work properly in FSX without turning off collisions because the aeroplane always crashes into the ground as it enters the runway. Presumably the FSX update will fix this.Tim

I should have mentioned that for the EGLL comparison I set the weather to "Gray and rainy" (which I think is one of the more graphically demanding pre-settings). Obviously both cards are a bit faster with clear skies.Tim

Thanks Tim.I know you were only comparing the two graphics cards but your FPS figures may help those of us who fly the 767 and have yet to upgrade to make a better spec PC.

Good: I hoped it would be helpful for upgraders and I rather suspected that quite a few people would also want to fly "framerate killers" like the 767.Tim