With many of us now taste testing the FSX demo, the question comes to mind of "how does my system compare to others with the same settings". To that end, I have produced a dynamic FSX demo benchmark that I hope others will try and report their results back to the forum.To run the benchmark, you will need to extract and copy the files from the attached FSDemo3.zip file into your Flight Simulator X Files in your My Documents directory. This contains the benchmark flight, weather and video files.You will also need to get a hold of a copy of FRAPS, which you can get from here http://www.fraps.com/download.php. I am using version 2.3.2, which can be used unregistered (ie. free) to do what we need to do now and hopefully the latest version allows this too.Set up your system as follows:1. Disable any secondary monitors you have such that you are only running the primary display. Besides FSX demo freezing when I use my PCI adaptor, I get incorrect FPS readings (double what is actual in fact) even if I have the secondary adaptor of my X800 running a monitor. You don't have to disable the hardware, just go into Display Properties and deselect "Extend my windows desktop onto this monitor" for all secondary monitors.2. Set your graphics card AA and AF settings in your graphics card control panel to Application Preference. These can be controlled within the FSX demo (well partially anyway) and provide much more consistent and comparable results between our systems.3. Shut down as many applications and system tray tasks as you can, without knobbling windows!4. Run FRAPs and ensure the following settings are made: Benchmarking hotkey - change to F1 (default F11 undesirably changes views in FSX and we don't need throttle control), check "stop benchmark automatically after ..." and type in 50 into the seconds box. Also check all options under the "save detailed benchmark statistics" section. Ensure at least one corner is selected for the Overlay (ie. not disabled) Minimise FRAPS and it should go to the system tray.5. Run the FSX demo, which should take you to the free flight page. Click on the settings tab, then the customise button to make the following changes for what I propose be the baseline benchmark for people with mid-high level systems:a. Graphics tab - initially set Global Settings to "Ultra High", then move the target frame rate to "unlimited", select "Ansiotropic" for filteriing and check the Anit-Aliasing checkbox. Also ensure the full screen resolution is set to 1024x768x32.b. Aircraft, Scenery, Weather and Traffic tabs - set global settings for each of these tabs to "Ultra High", then click on OK.c. Your display settings should now show as "customised" for graphics and "Ultra High" for everything else.6. Go to the free flight tab and click on load, select the FSDemo3 flight and click "Fly Now" to start loading the benchmark flight. Once loaded, you should see the FRAPS counter in what ever corner you picked for the FRAPS overlay corner option. Also ensure you are in full screen mode (again for consistency and comparability of results). Hit shift-Z twice to see the FSX frame rate counter and make sure that it is indicating approximately the same as what FRAPS is. If not just report any discrepancy when you post your results.7. On the FSX menu select options->flight video then double-click on FSDemo3. The video will start and you will have about 10 seconds before the benchmark starts to get everything looking right. By looking right I mean that you are looking straight ahead in the VC view (hit F9 if it is not) with no significant texture problems. If your textures are black/blue/purple as some have been experiencing, hit S then shift-S to toggle views, which clears the problem for my black screen issue in full screen. Also hold down ALT for a second or two to make the FSX menu disappear.8. Get ready to hit F1 to start the benchmark as soon as the left side of the windshield frame of the ultralight touches the bottom right corner of the painted runway number 0. I have attached a screen point of the exact point to hit F1. You will know you have triggered FRAPS into action when the overlay numbers quickly change colour then disappear entirely for 50 seconds while the plane flys.9. Watch the recorded flight and don't be too critical of my flying skills - the aim was to maximise eye candy, AI coverage and weather within a 50 second window! Not the recording will look jerky as it runs as a result of the recording rate of 1/4 sec, but it does not seem to impact FPS as you might think it would.10. After 50 seconds, FRAPS will signal the end of the benchmark by quickly flashing an FPS number with a dark background in your selected overlay corner. Again, a screen shot is attached of approximately where this should occur. Don't worry if you miss it because it has all been saved to file.11. Repeat steps 7 through 10 a couple of times to make sure you are getting a consistent FPS reading at the end.12. Minimise FSX and go into windows explorer to your C;FRAPSbenchmarks directory, where you should see a series of .CSV files containing the results of each run you did. The two useful files are the fps.csv (gives you and FPS reading for every second) and the minmaxavg.csv which gives you - you guessed it - min, max and average FPS for the 50 second run.13. Report back here with your results, indicating the following minimum information (if not readily evident in your sig): FSX settings, CPU & speed, RAM & speed, video card & speed, sound card, and average FPS for the run.If you have time, also report your results for the following different configurations to use at Step 5:a. Medium High - Set Graphics, Aircraft, Scenery, Weather and Traffic global settings to "Medium High". Go back to Graphics tab and set target frame rate to unlimited.b. High - Set Graphics, Aircraft, Scenery, Weather and Traffic global settings to "High". Go back to Graphics tab and set target frame rate to unlimited.c. FSXtreme (aka "Dr. it hurts when I ...") - Set Graphics, Aircraft, Scenery, Weather and Traffic global settings to "Ultra High". Go back to Graphics tab and set target frame rate to unlimited, Ansiotropic filtering on, antialiasing on, light bloom checked. On the scenery tab, move all sliders to the far right and check all checkboxes. On the weather tab, move the cloud draw distance slider to the far right and set thermal visibility to natural. On the traffic tab, move all sliders to the far right.Resuts for my main FS system are as follows:Athlon 64 3000+ @ 2.35GHz (~3700+), 1G DDR 3200 RAM, ATI X800XT AIW at stock 500/500, onboard AC97 sound:Medium High - 31.1 FPSHigh - 16.0 FPSUltra High - 13.5 FPSFSXtreme - 5.4 FPS I've showed you mine, now you show me yours!Gary

Hi Gary, it's definitely interesting to make such a comparison and I don't want to be a party pooper, but IMHO we need to consider a couple of details. First, there's Mike's (tdragger) comment in todays blog about AMD CPUs: >>- Some people are complaining about the default graphic settings. There is a bug where AMD Athlon PCs get down-graded because of the way they report their CPU speed. AMD users check your settings.<< Secondly, we're still looking at a pre-release demo. At this stage, I fear such a benchmark will lead to comparing apples with oranges and I would be very weary to draw any purchasing conclusions right now. According to SimHQ's AMD vs C2D comparisons - that is if I got it right? :-) - AMD64 CPUs lag ca 15% behind their Intel counterparts with FS9. Hence, anything way beyond such levels with FSX might have to be taken with a grain of salt? Again, I certainly don't want to spoil your fun or discourage you, but IMHO it's better if we keep some of these things in mind. Of course, comments, ideas and suggestions about the required mods concerning AMD chips are more than welcome. :-) Thanks and kind regards Jaap

>>First, there's Mike's (tdragger) comment in todays blog about AMD CPUs: Feel free to correct me if I am wrong, but I believe he was merely saying that the inital settings are downgraded from an optimum level for AMD processors, but that adjusting those settings manually is not an issue. It's not a performance issue for AMD-folks, just an issue with the default settings.I think is a good time to run benchmarks. It allows us to run benchmarks using a much stronger comparison than we can achieve with the released product. When FSX comes out, people will be mucking with the config files, splicing in FS9 add-ons, etc. At that point, benchmark comparions are meaningless.It would be nice to know how my older system compares to others, so I hope that this thread gets some input (and I will add my results).

>>It allows us to run benchmarks using a much stronger comparison than we can achieve with the released product. When FSX comes out, people will be mucking with the config files, splicing in FS9 add-ons, etc. At that point, benchmark comparions are meaningless.<< That certainly is a good point, Jimmy. I hope you see my main point to be weary if we see differences beyond normal? IMHO display adapters are highly important this time around. @ 1024x, my guess is a 9600pro will probably not deliver the same framerates as a X800. For even better comparisons we might have to reduce the resolution even further to 800x for example? HST, it's excellent Gary took that into consideration as well as possible in his HowTo and proposed a fairly low res to start with. :-) Cheers and kind regards Jaap

As stated by Jimmy, FSX may initial detect the wrong optimal settings for AMD CPUs, but this benchmarking exercise overrides all those settings anyway.I do agree that we should take the performance results of this pre-release demo with a big grain of salt, but IMHO it gives a reasonable idea of what to expect in two months and particularly lets someone know if they are in the ball park FSX performance wise compared to somebody elses similar system. Anyway, participartion is obviously voluntary and I think we should all promise not to give ACES too much grief over these preliminary performance figures.Speaking of performance figures, I ran the benchmark on my wife's computer to see how much impact single versus dual core has on FSX's performance. I did this in another thread, but that was a little less scientific in that I actually changed CPU's out of the same system (A64 and A64 X2), and they weren't running at exactly the same speed (2.35 versus 2.4 GHz).This time, I found a way to disable one of the cores using the boot.ini file (put /ONECPU at the end of the last line which starts with "multi..." and reboot, take it out and reboot when you are finished testing) and ran a true single versus dual comparison. My results were as follows:System - Dell Inspiron E1505, Core Duo 1.66GHz (not to be confused with Core 2 Duo), X1400 256M card with 128M of that shared, onboard sound.Medium High (single, dual, % boost) - 18.5, 21.1, 14%High - 10.5, 12.0, 14%Ultra High - 8.0, 8.4, 5%FSXtreme - 5.6, 6.0, 7%Of particular note are the minor boost to performance that dual core actually gave, the diminishing return of performance advantage dual core gives as setting levels are increased, and surprisingly that at max setting (FSXtreme) this system pulled slightly better numbers than my A64 rig did, despite it being beaten soundly in all other complexity settings.Gary

I was expecting to see a bit of a difference with video card overclocking this FS version around, but a a few quick runs using the Ultra High setting with my video card at stock 500/500 and overclocked to 550/550 revealed no discernable difference in the benchmark results. Even stranger since the Ultra High setting I have defined also includes in-game AA and AF enabled.Gary

Yeah, I saw that single vs dual-core test you ran. Very interesting. Every little bit is welcome with our favourite app, isn't it? :-) We probably cross-posted, Gary. Not knowing whether we can go to even lower resolution settings - my FlyTendo is running FS9 - wouldn't we maybe get slightly better results if we dropped the resolution even further? I.e. 800x? Just an idea. :-) Cheers and kind regards Jaap

Dang, we cross-posted again! :-) :-)

Jaap,I picked 1024x768 because it is a mid range resolution that should favour neither CPU or GPU and hence give a more balanced system assessment. By all means, test at higher or lower resolutions but just make sure you clearly note it in the results you post.Gary

And again! I'm done typing for a while :-lol Gary

Ok,here are my results. Is it correct that the tests for medium high and high are without Anisotropic and AntiAliasing checked in FSX?My system:AMD 4600 X2 DualCore @ 2640MHz, 2GB DDR 3200 RAM, Gainward Bliss 7800GS 512MB @ 500/700, Audigy 2ZSMedium High - 35.4 FPSHigh - 20.1 FPSUltra High - 18.0 FPSFSXtreme - 8.0 FPS

Yes that's correct. No AA / AF for the two lowest quality settings. Thanks for posting your results.Gary