Gary, just thought of something! Are you using traffic, and specifically default traffic? If so if you are near a major airport your results can be at least initially skewed because of the active jet ways which are activated by default airline traffic and any AI aircraft setup to use them. Since all the jet ways are being deployed at the same time until they have fully deployed, it can have a significant effect on performance. Once they are though, performance goes back to normal!! Most third party AI aircraft do not have the exit section of the aircraft.cfg file setup this way so they are unaffected by this.

Tom,Yes the standard test uses global high defaults, which set 15% traffic. As I did my test on a pristing FSX install, default AI were used. The test does start out about 15nm North of KSEA and it take about 2.5 mins for that airport to come into view, so maybe the jetways have done their thing by that time. I have now loaded UT-X AI traffic, so I shall see if there is any difference.FYI everyone, the beta has gone well so I'll be popping the file set in the AVSIM library either today or tomorrow for all to endure as they see fit :-)Gary

Just to add re traffic. Even if you have installed lots of new AI traffic, either the installer or you hopefully renamed the default TrafficAircraft.bgl to something with a different extension (eg. UTX changes the extension to .utb). If you want to return to default AI at any time, you just have to change the TrafficAircraft.xxx extension back to .bgl and change the extension of the newer traffic bgl files to something else.Also, you only need to go to these lengths if you want a valid comparison with another system. If you are just doing before and after hardware/settings changes on your own system, just leave everything as you currently use it. Afterall, you want to see what improvement or otherwise happened for your own FS world.Gary

Hey Marten,Can you please do a run of this benchmark on a pristine FSX SP2 installation on your new system. It will be really valuable to the FS community to find out just how spanky your next gen system at stock and overclocked speeds compares against a rig like mine, which is a pretty staple new system (minus the overclocking).I should have the files up in the library today or tomorrow, but if you are ready before then (or I don't make my own deadline), send me a PM with your email address and I will set you up with the files.Gary

Interesting Pierre, but what I see on my end too.... as a tip, try to keep the values on a graph of the same scale.... I mean on the Y-axis.... bottom only goes to 30...

It seemed logical to test again with FPS limited at 18.Here are the results. (with graph on same scale. tks ryan)Legend: Lim24 means FPS limited to 24; UNL means FPS unlimited and Lim18 means FPS has been limited to 18http://forums.avsim.net/user_files/181141.jpg1. This pretty much kills the theory. Doesn't it ?I understood that you would limit your fps in order for the 3d engine to better react to other processes: like loading textures, reduce stutters, reduce/eliminate blurries!!.....2. Honestly, the graph doesn't say more than FPS readings. It doesn't give hard numbers on texture loading or blurries.3. Speaking of texture loading, I didn't notice any change in texture loading during the tests. I made visual references for that. 4. Didn't notice anything with blurries either. I don't have them. Picture remained sharp during the tests at either FPS setting.5. Also, you should consider the hardware used for the tests: a. resolution: 1600x1200x32 60hzb. monitor: Viewsonic VP2030b (20") DVI-Dc. AA & AF application controlled (as per instructions)d. enhance application settinge. gamma correctionf. AA & Aniso on in FSXg. CPU E6850 3.0ghz (stock speed)h. MB Asus A5pn32e-sli (bios1205)i. RAM 2gig Corsair 800 (stock)j. FSB 1333k. Video eVGA 8800GTS 640meg (stock speeds)l. video driver 163.71m. Audio onboard AD1988n. FSX SP1o. UTX USA_CANp. GEXq. TREEXr. MyTFCX5.1as. Windows XP SP2, DX9 latest updatest. commit charge 188/3939. 26 services runningu. no FSX tweaksv. FSX bucket 7 proc speed 8379x. screen mode Full - menu bar offThis is based on Gary's FSX benchmark. Thanks Gary.Pierre

>Excellent work. I will publish fps rates for this system in>the coming weeks: >>QuadExtreme 45nm 9650 CPU @ 4 MHz and 5 MHz / Asus P5E3 />DDR3-1600 / Asus EAH3870 DX10.1 graphics / Raptor HDs/ >>I am in the process of installing everything, then run a>burn-in for a week, then proceed to install Vista 64 and FSX.>>>First tests with the system in a slightly different test>configuration with a different graphics card show more than>double the rates of the Q6600 reported above. That's great! I would like to see those results too, if you can test it!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

Having spent the last few days loading up all my addons for FS9 and FSX, I finally got to use FSX for something other than benchmarking - ala an online flight on VATSIM. As I was cruising along however, I thought I might do some performance data captures of FSX + addons in action, just to see what a "real" load is on the sim in actual use.For background, the flight I am reporting on below is not my Seattle benchmark run, rather I am on a flight from CYYT (St Johns, Canada) to BGGH (Nuuk, Greenland) in a Lockheed Superconstellation. This is the third last leg (of about 280!) of an RTW heavy iron tour I am doing with Air Source, and I decided to hook up on VATSIM to do the flight (A connie really confuses ATC when they see it on their scope!). Location at the time was out over open water heading North to the YMH VOR, about 80nm out.In addition to FSX, I had the following addons running:ASX - for weather of course!FSCommander - moving map, flight planner and for monitoring online flight progress on VATSIMTrackIR - head controlled panningFSBuild2 - used to create the flight plan originally, but left openFSInn - connects me to VATSIMServerInfo - for monitoring online flight progress on VATSIMActiveAirSource - the online ACARS system we use at AirSourceResults of a 5 minute log of CPU time for each FSX-related application are graphed as follows:http://forums.avsim.net/user_files/181142.jpgNote, the big spikes for both FSCommander and ServerInfo are when VATSIM data is redownloaded and updated on the screen.http://forums.avsim.net/user_files/181143.jpgIn summary, each of the applications averaged the following average CPU use (100% = 1 core fully utilised):FSX - 189%ASX - 7.6%FSCommander - 3.7%TrackIR - 2.4%FSBuild2 - 0.1%FSInn - 5.4%ServerInfo - 7.4%ActiveAirSource - 0.2%Total - 216%Translated, this mean that my system kept 2.2 cores on average entertained over this five minute window. Compare this to the FSXBench runs I did above, where they used about a 1.5 core average. The extra CPU use is not only from the addons themselves, but also FSX as it takes time to communicate with them (it alone goes up from 1.5 cores to 1.9 cores).So it would seem for this run at least, that two cores is just below optimal (note you normally wouldn't have FSCommander and ServerInfo both displaying VATSIM flight data, well unless you are me ;-)), and four cores is more than enough. Hey, Intel should think about putting out a triple core processor - the theory that three is best worked out quite well for the surfing industry at least :-lolGary

Gary, you are the benchmark king!What tool did you use to capture CPU useage per app?

I used the Vista 64 inbuilt Reliability and Performance Monitor, and I believe XP (well at least XP Pro) has the same performance logging capability. You can either set up the Performance Monitor live to view, or create a data collector set to log, % CPU load by individual process name. Excel analysis and graphing trickery do the rest! Gary

Gary,Are you noticing any FPS or other performance impact on FSX with all those addons running or are the remaining cores picking up all the additional load?Glad to hear you are getting some flying time in.Ted

ASX still causes a few stutters during initial hookup with FSX, but that is to be expected given the high amount of comms it has with FSX during initial weather loading. After that, it's all smooth.With the dual core I had, I would notice stutters and FPS dips during ServerInfo's VATSIM data update and redraw every few minutes, but now with quaddage that extra 100% CPU load is gobbled by one of the spare cores, those stutters have gone. Re FPS impact, it is too hard to tell, as FPS float about all over the place during normal flying anyway. What I am seeing now that I have left the cityscape of this benchmark is that I can fly with most sliders right and have FPS in the 40-80 range. In fact, flying over desolate Greenland, I see triple digit FPS quite often. :-)Gary

For those that are interested, my benchmark package has now been uploaded to the AVSIM library. Here is a quick link: http://library.avsim.net/search.php?Search...xutil&Go=SearchWhilst this test has many uses, it will be particularly useful to see how the new generation Penryn and Phenom systems perform with FSX, particularly compared extant line of multi-core systems as I have tested above. Accordingly, I request that those purchasing one of these new systems take the time to do at least a standard test baseline run (as defined in the package) of either an FSX SP1 or SP2 pristine installation and report back here with the results.Gary

The way I've noticed, is that if you lock you FPS and can constantly attain the value you set, you have less blurriesBut if you lock at say, 25, and only get 25fps 20% of the time, you will still get blurriesUnlimited also gives me more blurries but better fps - not necessarily smoother, because I get stutters if I use unlimited....

Excellent! Thanks for all of your hard work! This will help folks like me who are contemplating hardware upgrades. Hopefully you will be able to add to this once new HW becomes available.RH