I have read a new discussion about agp aperture size and always had it on 128 mb.Just for the heck of it, I tried putting it all the way down to 32 mb.Guess what?fps increased by +/- 5 and ALL games, especially fs2004 and LOMAC run considerably more smooth.Had a discission with my brother who is an IT specialist, he sad it is impossible because all the agp aperture setting does is setting the amount of phisical memory to use when the video card runs out.Then he tried it anyway, and guess what - he was simply amazed!!!(Told me maybe a problem with reverse adressing in directX9.0c or windowsXP - or something - ).Actually, here is what i think, when you set the aperture size lower you are forcing the video card to USE its memory, and as we all know its memory is mostly 3 to 4 times faster than our phisical ddr pc3200 ram.Remember though, do not set it too low, because you will disable the agp port. But for most new mobo's you cant set it lower than 32mb, which is a good setting.Please give your responses, as to what you've found...Athlon XP 3200+Asus a7n8x-x512 mb PC3200 (400Mhz)Ati Radeon 9800 PROMaxtor 40Gb (7200rpm)Catalyst 4.8DirectX 9.0cWindows XP sp1

Hi,Interesting! I will definitely try when I get home from work. Have a similar setup to yours so it should be a good comparison. Thanks for the tip!Athlon XP 3000+Asus a7n8x Deluxe1Gb PC3200Ati Radeon 9600XTCatalyst 4.4DirectX 9.0cWin XP Pro SP1Regards,Jurehttp://www.globecargo.org/images/VAA_Captain.jpg

Hi,By the way - you might want to read this:http://www.videocard-forum.com/ati/AGP_Ape...ize_184600.htmland look where he describes agp ap with simulations, i.e. LOMACOverall I would just keep it set at 32MB and change it to 64MB or 128MB when playing a first person shooter.However, I am a great flight sim. enthousiast, so for now i leave it at 32MBFinally, if an application needs more memory, it will use your phisical ram anyway, no matter how you have your aperture set.And if it doesn't ("-- give me a call, i'll fly with you--") and you do get problems - just bring it back up or down, it's not like your computer is gonna die from changing these settings.Greetings - Anthony

Anthony,indeed it works! I gained about 2FPS in a pretty dense scenery, from 25 to 27. AGP 8x is still on and so far, I haven't seen any bad side effects. Thanks again! :)Jure

I don

This topic keeps returning back and I think it is good ...I have no clue why but some people on this forum (and otherwise quite savy in computer matters) keep advocating setting the AGP higher if you have more system memory. And really AGP apparently should be set based on how much memory is on your video card, not your RAM. The more memory your card has the less AGP you need.Just for an experiment I set AGP to 32 (from 256) and gained about 1% in final score running 3Dmark03. Not much but as you can see I did not loose any performance. And I am yet to try flight simulation but I believe folks who say it results in some perf. gain.Michael J.WinXP-Home,AMD64 3500+,Abit AV8, Radeon X800 Pro,WD 36GB Raptor,1 GB PC3200 http://www.reality-xp.com/community/nr/rsc/rxp-higher.jpg

Good to be of help, Jure.Indeed Michael, you are right. However keep in mind that the low setting applies to flightsims only. As i have stated and tested, when playing a first person shooter or "room" game, you need to set it a little higher; 64mb is sufficiant.cheers,anthony

But then if everyone start advocating set it as low as possible we will have the same problem. The fact is that it most often don

>The fact is that it>most often don

But then I have seen more that recommends 256 in aperture size which just shows that it

I've been experimenting before with aperture size and I saw no difference - at least not noticable. The stutters, those micro stutters when ATC speaks have always been there and I couldn't say the textures loaded quicker, either. But today I saved a situation before setting the AGP aperture to 32 and watched the fps for quite a while to get the mean value. I got 2 fps plus, no doubt. :)After that I cleaned the remnants of nVidia drivers, but there were none(!).Last, I upgraded from 4.4 CATs (were the best, 7 and 8 gave me stutters) to 4.9 and... lo and behold, I got another 3 solid fps! Am I happy? You bet! :-lolThat's from 25-30, so that's QUITE a lot for an hour's work.Jurehttp://www.globecargo.org/images/VAA_Captain.jpg

Again, glad to be of help.However, to the contrary to your findings on cat 4.9, i found that 4.8 (on my system) is a LITTLE better. But after reading endless discussions which is better - i came to the conclusion that its kind of a matter of suggestion.They are both good and should of been as from 3.9 (Which was the one I used for months).Furthermore, it was never my intention to generalize flightsims. But they mostly work in the same way and ARE generally very cpu and ram dependant. Whereas room based games with special textures like (enviromental mapped) bump mapping, T&L, etc, are more gfx card dependable. That is why ATI scores high with AA or AF enabled, because this is (ofcourse) done by the card.Ofcourse Oqvist has a point and it is possible that on his system il2 FB runs (better) with 128mb agp ap - however i installed it, left the setting of 32mb and noticed that it runs better than before i had it installed - which was a couple of months ago.Having said that, my brother claims that he is absolutely happy with the setting. And he is an absolute battlefield / half-life / doom3 player. He says that with doom3 he gained around 5 fps and an overall smooth play - but that involved fare more tweakings than i can write here.It is ofcourse the goal of all us flightsim lovers to have fs2004 running as smooth as possible with all gfx settings at max and all addons installed, and to help each other getting there weither you have a low, mid or high end system. If you have any questions regarding hardware setup, please ask, as i am happy to answer.Pleasant flying,Anthony

The AGP setting is not a one size fits all...Different systems and chipsets and even BIOS revisions or even drivers will behave differently to AGP settings even though one might be using the same CPU, video card and system memory. If you are able to test this on different machines running the same benchmarks you find this to be so.The best advice is to test and benchmark for the best setting that your system likes and as observed by others smooth stutter free performance is as or is more important than any 5% FPS gain that might show up in an average of >good< benchmarking, and folks if you are discussing FPS than you HAVE to benchmark, general observations are completely untrustworthy and not worth considering.For the system you see specked out below right now a 512mb AGP setting has given consistent FPS increase of 0.5 FPS on a average of 18FPS using 1600x1200x32 2QxAA and 8xAF in FS9 under dense scenery settings with 38m mesh and my custom scenery (1400 to 2800 custom textures -dusk/dawn transition) while remaining completely stutter free while flying CS727 in VC view.A new 256mb 6800GT will soon fill the slot for this system as well as some quicker system memory so it will have to benchmarked again, the only thing in the display settings that will change is that AA will increase to 4xAA.The theory that newer 256mb cards or even upcoming 512mb cards need less of an aperture is lost due to the fact that the user will fill the frame buffer even more with the same game through use of higher screen resolution, higher AA and AF settings than was used on their previous card that had less of a frame buffer and the need for the AGP aperture may still be required as before.Regardless of any claims by anyone anywhere AGP aperture settings should always be tested and applied to the application and on the actual hardware that it will be running on.

>The AGP setting is not a one size fits all...>>Different systems and chipsets and even BIOS revisions or even>drivers will behave differently to AGP settings even though>one might be using the same CPU, video card and system memory.>>>If you are able to test this on different machines running the>same benchmarks you find this to be so.>>The best advice is to test and benchmark for the best setting>that your system likes and as observed by others smooth>stutter free performance is as or is more important than any>5% FPS gain that might show up in an average of >good<>benchmarking, and folks if you are discussing FPS than you>HAVE to benchmark, general observations are completely>untrustworthy and not worth considering.>>For the system you see specked out below right now a 512mb AGP>setting has given consistent FPS increase of 0.5 FPS on a>average of 18FPS using 1600x1200x32 2QxAA and 8xAF in FS9>under dense scenery settings with 38m mesh and my custom>scenery (1400 to 2800 custom textures -dusk/dawn transition)>while remaining completely stutter free while flying CS727 in>VC view.>>A new 256mb 6800GT will soon fill the slot for this system as>well as some quicker system memory so it will have to>benchmarked again, the only thing in the display settings that>will change is that AA will increase to 4xAA.>>The theory that newer 256mb cards or even upcoming 512mb cards>need less of an aperture is lost due to the fact that the user>will fill the frame buffer even more with the same game>through use of higher screen resolution, higher AA and AF>settings than was used on their previous card that had less of>a frame buffer and the need for the AGP aperture may still be>required as before.>>Regardless of any claims by anyone anywhere AGP aperture>settings should always be tested and applied to the>application and on the actual hardware that it will be running>on.>>>QFTGreg

I wonder if one could set up the PMDG FMC to create a flight so you go through the same exact path every time for a reliable benchmark?