I am not surprised. Wait till you adjust cloud settings to max and ground shadows, it's almost like a new sim.Glad it did it for you too. Those 8800 video cards are super.Pierre

I am very, very surprised. I did not expect it. I have all my setting maxed, including ground shadows and clouds. RH

I too am very VERY surprised. In my experience, using 4X/8X AA/AF as my standard setting with FS9, I got the biggest boost going from a GF3 Ti200 to a 9800 Pro, but video cards thereafter (6800, X800XT, 7900 and 8800) have made neglible impact to low-end FPS scenarios. If the improvement is as great as you say, then I reckon your new video card installation flushed out a bad driver install that was making your system run suboptimally in the first place.Gary

It kind of reminds me of the FS98 days when I switched out an old Voodoo card to the Nvidia TNT card.The last time I got results half this good was going from my Voodoo 3 2000 to the GeForce4 Ti 4200 using FS2002. Using FS2004 and when moving from the 4200 to ATI 9800 Pro, I did not get any increase in frames per second, but I was able to utilize some AA and AF without loss of frames-per-second and the quality looked better. Back then, I used an Athlon XP Barton (overclocked 2.436 Ghz).My scenario was the following (using the Barton 9800 combo: Sitting at the default Seattle Int runway in the default C-182. All sliders maxed (1600x1200) with the exception of 30 miles visability, partly cloudly. I used X2 AA and X4 AF - I believe. I just know those two settings were not maxed.My frames per second ranged from 11.8 - 13.5When I switched to an ASRock MB with an AMD Athlon 3800+, using the 9800 Pro, I obtained the following frames per second range with the same AA and AF settings: 17 - 22.4 Also, in FS2004, when I dumped my 9800 Pro for the PNY Nvidia 7900 using running an AMD Athlon 3800+, I did not get any increase in frames per second, however, I was able to turn up AA and AF all the way with no loss in frames-per-second. Also, there was a quality improvment as well. I was still 17 - 22.4I then build a new system around my current Athlon FX 57. My frames-per-second went to the following: 23 - 27 with the 7900. I installed the latest drivers from the Nvidia site, so there should have been no driver issues.When I overclocked the the 7900, I saw no improvement in frames-per-second. That was what led me to believe that FS2004 was CPU limited. I figured that I had maxed out the videocard end of it.However, I also then overclocked my FX 57 from 2.8 Ghz to 3.0 Ghz and also saw no improvement.Since that time, I had started putting add-ons like Level D, Flight Enviroment, Ground Enviroment Pro, Active Sky, and Radar Contact. I usually max out all options on those programs. The only times I had problems were with heavy cloud coverage in placed like JFK and that. My frames would occassionally drop to the 9-10 range. That's when I inquired into upgrading to the Blackbox XP 6400+ and then someone suggested trying an 8800. I was very skeptical (sp?). I then read the post of some others who had upgraded to the 8800 with great success. I then figured what could I lose? If it did not work, I could always run it back to Best Buy for a refund. I placed it in and with the old scenaro, I got frame rates in the the following range: 58 - 64!! and higher once airborn. I about fell out of my chair. But then figured that my maxed out FS204 setting must have reset. They had not; they were still maxed.Now in places with heavy clouds and complex scenery, I am getting frames in 18-24 range, never lower that 14 to 16 (and that was with a complex ATL scenery I downloaded from AVSIM. Before with the 9700, I was getting as low as 7-9. Since upgrading with this 8800, I really believe we were wrong assuming that FS2004 was entirely CPU bound. I just don't think the right video cards were out. I am curious as to why I got such an increase you did not.

Very interesting numbers. One point to note is that we're talking about 3 generation of video cards here: from a GF3 TI200 to a 8800GTS.To put things into some kind of perspective I've built this small comparison of the video cards mentionned in the thread.http://forums.avsim.net/user_files/181785.jpgAlthough the table does not tell us the FPS will double in Flight Simulator, it does tell us about the "huge" increase in performance these cards have gone through the years. I believe the memory bandwidth is a good overall indicator of the performance of a video card.Also, I would venture to say that historically FS performance was mostly limited by the CPU and that is starting to change towards CPU and GPU limited. It is a balance between the two processors.Reference: www.gpureview.com (excellent site to compare GPU's)Pierre

I like that graph. I have just purchased Ultimate Traffic for FS9. I did not because of the extra FPS that the 8800 allowed me when compared to the 7900GS. I have more to burn, that is. Also, with the 8800GTS, I can use the heaviest form of clouds in Flight Enviroment and Active Sky, without much loss in frame rates (that is, again, when compared to the 9700. Overall, my frames have improved from 50% to 300% using the 8800 when compared to the 7900! From my results, I'd have to agree that memory bandwidth is most highly coorelated with FPS in FS2004. That coorelation probably works only with the top end CPUs of today (including my overclocked FX57 @3.0Ghz).RH

I have posted the following chart before, but I will do so again to illustrate just how CPU bound FS9 is:http://forums.avsim.net/user_files/181792.jpgAs you can see, doubling the resolution (normal = 1680x1050, spanned = 3360x1050) and cranking up AA/AF from nothing to full blast makes diddly squat difference to the lowest FPS acheived. And even under this max load, the GPU is barely cracking a sweat, with temps just a little above idle, and meanwhile the CPU is working its little heart out. This clearly shows that when FPS going gets tough, the CPU is determining just how low that bar will go.I am very happy that you guys are getting such good results moving up to an 8800, but in my experience on my current system as well a multitude of other people's systems I have set up with FS over the years, the GPU is not the droid you are looking for!Gary

I would not make it up. I am as serious a flight simmer as most on this board. If I made it up, I would be influencing others to waste their money (or time if they had to return it) on an 8800. I would not do that, because I utilize this board a lot for acurate information. I would not want others to mislead me. I would not want to influence others to fill up the board with innaccurate information so that it would be useless to others. The above results explain perfectly the transition from my old ATI 9800 Pro to the 7900GS when I saw no frames-per-second increase (but then again showed no decrease in performance when I turned up the AA and AF, like I did with the old ATL 9800 Pro).Maybe it was a bad driver on my 7900GS, but something had to make it give me 30% to almost 300% framerate improvement when inserting my 8800. However, that's an issue I definitely want to explore. In blue skies with minimal scenery, I get almost a 300% improvement. In a complex airport with lots of clouds and traffic (and when I expand the viewing area far enough) my improvement is only like 25% to 30%.Again, I'd like to explore this. What CPU did you use for the above benchmark? RH

Gary, I am not questioning the cpu load in FS9. And this is what your graphic is showing. That's fine. What your graph is not showing is the performance increase you get when you SWITCH from a dx9 card to a dx10 one. This is what the discussion is all about (I think) : The performance increase in FS9 when you switch video cards (from a DX9 to a DX10 one).The only real answer to this would be to bench a system with one monitor, same cpu, same memory but different video card: let's say a 7900 vs 8800. (unfortunately, I don't have the 7900 video card)Pierre

If you get really bored over the holidays, you could re-insert the 7900 and report on your findings... no software changes.In the meantime, enjoy !

Pierre,>The only real answer to this would be to bench a system with>one monitor, same cpu, same memory but different video card:>let's say a 7900 vs 8800. (unfortunately, I don't have the>7900 video card)>PierreAs reported earlier in this thread, I have done exactly this but with FSX instead of FS9 per the following graph. Ignore the spikes (later turned out to be an audio driver issue), but you can see that going from a 7900 to an 8800 only made about a 5% difference overall. And FSX is much harder on the graphics card than any previous iteration of FS is, so the gains in FS9 would be neglible.http://forums.avsim.net/user_files/181809.jpgGary

I am not saying you are making it up, rather that something else must be afoot given that my, and many others, experience is somewhat different to yours.I don't dispute that blue sky / simple scenery FPS goes up by such a huge amount, as I too have experienced mega FPS in such areas. But seriously, who cares what your top end FPS is when it is low end FPS that you seek to improve? And in any complex, FPS draining FS scenario I have been able to conjure up over the years, CPU speed has been proven to be 90% determinent of what FPS you are seeing, with memory and video card sharing the next 9% between them, and the sound card trailing in at 1%.The CPU I used with the above benchmark was a C2D dual E6400 at 3.2GHz. With the default FS9 flight (ie. Cessna @ KSEA), max sliders, AA/AF at anything 4X/8X or lower, and using either the 7900 or the 8800, I get about 45 FPS.Gary

All you can say for sure is that until recently, most reports from simmers supported the hypothesis that FS9 is primarily CPU dependent. Now we have evidence that that hypothesis no longer holds. FS9 does not appear to be as CPU dependent as once thought. However, I am not sure what aspects of my CPU/Videocard combi (and reports from others like me) created the sudden (and unexpected) improvement.Before this past weekend, I was squarely in your camp. I like you graphs and enjoyed reading over your hard work concerning hardware and FSX on your other posts.RH

Robbie, no disrespect, it is certainly interesting to get your feedback with this upgrade, but I for one am still having trouble understanding all this... In the initial euphoria you reported "double the framerates in all situations", then a day later 25% improvement... I'd still like to see some benchmarks that show the framerates in identical situations when the system is under heavy load. 25% improvement is good, but once you are in the single digits, 25% better is still unsatisfactory.

You want some great frame rates? I have finally found them!I upgraded from an AMD4800+ 7900GT to a Conroe E6850 with 8800GTI am still shocked how much a performance boost this is.For example, in situations where i would be getting 15-23 FPS ( for example the FPS hog PDMG 737-700) , now i am steady around 30-35 . I tested it with the CanUK CYYC with 100% AI traffic, every setting to the right, 2 layers of scattered cumulous/stratus ( default clouds). I am also running 16X AF and 8X supersampling AA.While landing on runway 16 was a stutter fest before, now the landing is as smooth as butter. Was like flying in a real aircraft. Simming in FS9 is now {" as real as it gets!)..A CPU upgrade ( especially a Core Duo/Quad) will give you a huge boost right there. A video card upgrade normally will not give you higher FPS, however when running heavy clouds and high AA settings, a faster card will help there...As for people saying they are having problem with Nvidia drivers, i am using 169.02 8800GT drivers with no issues so far. I had a bit of an issue with shimmering textures, however using Supersampling instead of Multisampling will reduce the effect.Plus a system like this will enable you to upgrade to FSX and get good performance too. You get the best of both worlds