AMD Dragon Platform

[Guest Surfstud31]

Since my original thread deteriorated into an ATI vs. Nvidia flamewar, I thought I would start a new thread to post my FSX testing results with the new AMD PC I built. To quickly recap, my AMD Dragon Platform consists of:Cooler Master Storm Scout caseASUS Crosshair III Formula moboAMD Phenom II X4 955 BE CPU OC'ed to 3.81 GHz8 GB Mushkin DDR3 PC-12800 RAM running at 1587MHzATI Diamond Radeon HD 4890 XOC 1 GB GDDR5 video cardWD VelociRaptor 300GB HDDCorsair 750-Watt Power SupplyFor this test, traffic settings are at 35% for all, aircraft graphic settings are ultra-high (cast shadows on self, etc), all little extras like detailed land tectures and ground scenery shadows are ON. Clouds are set to complex with 80 mile radius. The test was done in DX9 using the following add-ons which were all active:FS Genesis mesh (v 2.0 USA), v 1.0 rest of worldActive Sky X SP4X Graphics SP4Ground Environment X - USA & Canada (GEX) v 1.08MegaScenery X HawaiiMegaScenery X PhoenixAerosoft PBY Catalina XFlight 1 Pilatus PC-12Shockwave P51D MustangThe first screenshot shows the graphic settings for my FSX. Second shot shows the Aerosoft PBY over Honolulu in MegaScenery X Hawaii with 53 FPS. Third shot is from the PBY VC over Honolulu with 51 FPS. My FPS while flying the PBY over Honolulu remained in the 50's. The third shot is the Flight 1 Pilatus PC-12 over Phoenix in MegaScenery X Phoenix with 56 FPS. The fourth shot is over Phoenix in same plane from VC at 46 FPS. My FPS while flying over Phoenix remained in the upper 40's thru 50's. The last two shots are the Shockwave P51D over Manhattan in very overcast skies. Both those shots are at about 21 FPS. My FPS over Manhattan in heavily overcast skies was in the lower 20's.I am once again pleased with the results. I have not even OC'ed my video card or tweaked my CPU and RAM OC settings yet. These are the most demanding flying scenarios I could find on my setup. I have no PMDG planes yet so could not test one. The planes I did pick, however, I think are a good representation of some of the more demanding add-on aircraft available. I may try some DX10 tests very soon but now I am going to go and enjoy my new creation! Here are the pics:SETTINGSPBY OVER HONOLULU MEGASCENERY XPBY OVER HONOLULU MEGASCENERY X VCFLIGHT1 PC-12 OVER PHOENIX MEGASCENERY XFLIGHT1 PC-12 OVER PHOENIX MEGASCENERY X VCSHOCKWAVE P51 OVER MANHATTANSCHOCKWAVE P51 OVER MANHATTAN VC

[tfm 170]

These look like impressive results IMHO: thanks very much for posting them as it is unusual to see much information about how well AMD's offerings perform with FSX.For those of us interested in exactly what's going on under the bonnet, would you mind posting your settings for antialiasing and anisotropic filtering in the Graphics tab in FSX? Are you overriding any of those settings through the ATI control panel? And what is your resolution?Many thanks,Tim

[Guest Surfstud31]

These look like impressive results IMHO: thanks very much for posting them as it is unusual to see much information about how well AMD's offerings perform with FSX.For those of us interested in exactly what's going on under the bonnet, would you mind posting your settings for antialiasing and anisotropic filtering in the Graphics tab in FSX? Are you overriding any of those settings through the ATI control panel? And what is your resolution?Many thanks,Tim

I am happy with the results as well Tim. I have both anisotropic filtering and anti-aliasing ON in FSX (see pic). In the Catalyst Control Center, I check the boxes that say 'Use application settings' for both AA and AF. I had read somewhere that it is best to let FSX control those functions. My resolution is 1680 x 1050 as that is max for my current monitor.I still have not really tweaked my overclocking capabilities. Right now I am running what the BIOS automatically OC'ed for me at 3.81Ghz on all four cores. I would like to try and do an extreme OC on one core (say 4.4Ghz) and then run FSX on that one core. There are as yet many possibilities for me to try and that is part of the fun.

[TechguyMaxC 516]

Why are you running 16-bit color depth?

[Guest Surfstud31]

Why are you running 16-bit color depth?

Ooops...my bad. Had accidentally hit the wrong option when I went into the settings screen to take the shot. This is the correct settings screenshot (thanks Max, might not have noticed that for a while!):

[Guest UlfB]

Mike,I would like to see your fps measured with FRAPS instead of FSX. The FSX fps isn't reliable and will normally fluctuate a lot.

[tfm 170]

Wow: light bloom on! I spent a few minutes with your settings. So far as I can tell your results are broadly the same as mine, for a fraction of the price. Very interesting - and slightly depressing.Tim(i7 975@4.4GHz, 6GB 1871MHz 7-7-7-18 1N, GTX285@702MHz).

[Guest djt01]

I honestly don

[ryanbatc 12,905]

It looks pretty good! I think the difference between your CPU and a Core i7 would be texture loading, normally I see very clear crisp textures with guys' and their OCed i7's Yours look pretty good but I can see some blurries hehe - to be picky

[Guest Surfstud31]

OK, downloaded FRAPS and will take some more shots tomorrow. Spent some more time tweaking the system today. Time to go to bed.I never said my system was better than an i7 system. I just decided to build an AMD platform to save money and post my experience for those interested.

[tfm 170]

OK, downloaded FRAPS and will take some more shots tomorrow. Spent some more time tweaking the system today. Time to go to bed.I never said my system was better than an i7 system. I just decided to build an AMD platform to save money and post my experience for those interested.

Exactly: it's not a race. Posts like yours are extremely useful for helping others to make informed decisions, IMHO.Tim

[Guest Surfstud31]

Here are some more screenhots now using FRAPS. That is a neat little program, never used it before. It settles down the huge variations that you see in the FSX FPS indicator and gives you a more stable average. These shots are in DX9 with my previous settings and in the Shockwave P51D Mustang add-on. The first is Manhattan at sunest in ASX with FRAPS showing 21 FPS average. The second is same shot from the VC with 22 FPS average. Last is the horrible weather shot over Manhattan and still holding 18 FPS according to FRAPS. All in all, I am pleased with the results.I took a look at what Mike_CFII_MEL has posted and we have different settings. For one, my cloud draw is on at least 80 miles, his was minimum. I also have water effects on 2 high, he was at least 2 settings below that. If I find the time to match our settings exactly, I will post some shots although my point here was to share my experience and not get into a contest with anyone. I do have a life.NYC SUNSETNYC VC SUNSETNYC BAD WEATHER

[Guest UlfB]

Thanks for the FRAPS fps Mike!Seems that the AMD gives some impressing fps for the bucks.

[tfm 170]

For the record, doing my best to duplicate the Bell helicopter flight over NY shown in Mike's earlier thread, I get about 15 fps where he gets about 27 fps. This is on an i7 975 @ 4410MHz with 6GB of 1866MHz RAM at 7-7-7-18 N1, with a GTX 285 factory overclocked to 702MHz. Ie: according to conventional wisdom, about as good as it gets - and at a pretty high price. The only differences in our settings that I can detect are that (i) my screenshot is in 1920x1020 resolution - but I have tried with Mike's resolution too, with no appreciable difference; and (ii) I'm using FEX clouds and water textures, but the texture resolutions are not enormous so I can't believe this is making all the difference.I have also tried with Nick_N's nHancer settings (which is what I usually use): no real difference.So unless something is badly wrong with my setup*, Mike's onto a complete winner with this kit.I think this is quite important news for FSX users.Tim* EDIT: Which is possible: I transferred my whole hard disk setup lock stock and barrel to my new i7 PC, without any fresh installations. Win7RC just seemed to install the new kit as if it were a new device. Bob Scott pointed out in an earlier post that this method might have prevented Windows from recognising the new kit properly. Maybe: but it's much faster than the old E8600-based PC and other benchmarks are what I would expect (~21000 in 3DMark 2006, ~9s for SuperPI 1M, etc), so I rather doubt it.

[Guest djt01]

For the record, doing my best to duplicate the Bell helicopter flight over NY shown in Mike's earlier thread, I get about 15 fps where he gets about 27 fps. This is on an i7 975 @ 4410MHz with 6GB of 1866MHz RAM at 7-7-7-18 N1, with a GTX 285 factory overclocked to 702MHz. Ie: according to conventional wisdom, about as good as it gets - and at a pretty high price. The only differences in our settings that I can detect are that (i) my screenshot is in 1920x1020 resolution - but I have tried with Mike's resolution too, with no appreciable difference; and (ii) I'm using FEX clouds and water textures, but the texture resolutions are not enormous so I can't believe this is making all the difference.I have also tried with Nick_N's nHancer settings (which is what I usually use): no real difference.So unless something is badly wrong with my setup*, Mike's onto a complete winner with this kit.I think this is quite important news for FSX users.Tim* EDIT: Which is possible: I transferred my whole hard disk setup lock stock and barrel to my new i7 PC, without any fresh installations. Win7RC just seemed to install the new kit as if it were a new device. Bob Scott pointed out in an earlier post that this method might have prevented Windows from recognising the new kit properly. Maybe: but it's much faster than the old E8600-based PC and other benchmarks are what I would expect (~21000 in 3DMark 2006, ~9s for SuperPI 1M, etc), so I rather doubt it.

If you