Interesting FSX performance comparison.

[Guest azflyboy]

http://www.extremetech.com/article2/0,1697,2059176,00.aspThis is a pretty stunning indicator of how hard FSX actually hits systems in terms of performance. The tested systems run about $7,000, and given that FSX delivered less one third of the performance of FEAR (which is famous for delivering poor performance) on those machines and never broke 20FPS, I'll be very interested to see what kind of performance gain we get from patches.

[Guest FxF3]

This has been posted before. A few guys with computer specs like that have already showed us that FSX does great on a high end computer there have been lots of photos showing FPS in the high 30's and mid 40's in the big cities and airports with those machines. Do a search.

[bdenley 0]

No Bruce. Those gradings are measured, according to the article, with tools supplied by the FSX development team. It would be intesting to see the FXS / FS2004 comparisons but one thing is clear: FXS brings the best systems money can buy to it's knees. Imgine how it will perform on your system!

[DadJokeCinema 145]

>No Bruce. Those gradings are measured, according to the>article, with tools supplied by the FSX development team. It>would be intesting to see the FXS / FS2004 comparisons but one>thing is clear: FXS brings the best systems money can buy to>it's knees. Imgine how it will perform on your system!I don't know, I have a $1200 computer that runs solidly at the mid-20's.Of course, I also had no issue with FEAR, so I can't imagine what it could be without looking at those results in detail. More than likely, if they are like most benchmarking sites, they are testing at rather high resolutions with lots of AA and AF, and all those extra passes can certainly bog down a system.I'm gonna take a look at the article and see where the departure between my performance and theirs is at.

[Guest azflyboy]

I agree that the framerates shown in the tests were lower than a lot of people (myself included) get on far cheaper systems, but these benchmarks are done with every possible graphics setting turned to high, no tweaks whatsoever, and at very high resolution, so it does show that FSX is much harder on PC's than most other games.

[tf51d 284]

>No Bruce. Those gradings are measured, according to the>article, with tools supplied by the FSX development team. It>would be intesting to see the FXS / FS2004 comparisons but one>thing is clear: FXS brings the best systems money can buy to>it's knees. Imgine how it will perform on your system!One I have a Gateway FX530XG, which was one of the systems tested in the above review, although I have an E6700 not the X6800 though. I can tell you I get much better results than stated in the review. One thing you don't know in the review is what settings were used, whether they took into account the default.xml bug, which has not been fixed. My guess is they have the settings high, which MS has already stated is not meant to be used with even todays high end cpu's. Those settings are meant for future even faster systems. The idea is to build in expandability to prolong product life. They have basically done this with every FS version to date. Even at these reduced settings, you can get higher detail than FS2004. If you want to see what kind of performance currently available high end systems can produce in FSX, check out my zipping around Manhattan video in the video forum. My settings were extremely dense scenery, dense (Though tweaked) autogen, texture resolution set to 1m mesh set to 19m, weather set to max and 100% traffic using a modified, though still very large traffic file with UT flightplans. You can also see a couple of my videos on youtube which also shows the performance I get in FSX. (The micro stutters are a result of the fraps recording and video reduction, the actual flight was very smooth with frame rates in the mid to upper 20's for these flights, measured with the on screen display from ATI Tray tools. Unfortunately it doesn't show on videos.)You can find them by searching tf51d on youtube.com

[Guest]

FS brings nothing to its knees. FS wastes system cycles and FS does *******************NOT************ use dual cores, virtual cpus etc. in effect, FS is the one of the few benchmarked games that runs on half the system. now that we've gotten this out of the way let's just relax and move on...

[Guest]

FSX is the only game that is "built for the future" that doesn't use current hardware to it's maximum potential, if at all.

[Murmur 3,941]

>FSX is the only game that is "built for the future" that>doesn't use current hardware to it's maximum potential, if at>all. If FSX is made for hardware 3 years in the future, does it mean we can pick it up and then pay $69.99 to Microsoft only on 2010? :(Marco

[FTD1949 0]

With all due respect it's not a stunning indicator. All of the tests were done at resolutions of 1920x1200, and 2560x1600. They are also running with 4xAA and 8xAF. If a person used 1280 x 1024 on their machine and then jumped it up to 1920 x 1200 or 2560 x 1600 they are essentially increasing the resolution by 76 % with each move up and of course it's going to cause a big hit. I really don't believe that any but the most extreme hard core user uses those screen resolutions. I don't have a $ 5000 machine and don't expect to have one in the near future. I achieve framerates on RW 34R at Seattle of 15 - 17. When airborne I avg close to my lock of 35 fps. My settings are as follows: Res. 1280 x 1024 x 32 (native res of my 19" LCD) Global Texture - High Advanced Animations checked Bilinear filtering AA checked High Res VC Landing lights checked LOD radius small Mesh Complexity 80 Mesh Resolution 5m Texture Resolution 1m Water effects 1/x Scenery Complexity Dense Autogen None Special Effects Medium Airline Traffic 30% General Aviation 10% Road,ship Leisure 0% Detailed Clouds Thermal Vis Natural Weather Changes Medium Winds aloft checked I have no problem with fluid flight AMD64 3500 Venice 2GB Corsair RAM evga 7900 GT KO nvodngov19147-[Guru3D.com]drivers SB Audigy2 Sound ASUS A8N Deluxe SLI CH Yolk and Rudders Track IR3 w/vector Radar Contact AS v6 I'm sure I can bring any machine to it's knees one way or another but to believe that those machines can't run FSX without bogging down is just naive.

[Paul J 349]

. . . and the processor future is multi-core, so the future will be Flight Sim ELEVEN. I've predicted this in a number of posts. FS 10 will have a short life - maybe 18 months at the most and MS will use everything they're learning right now, and roll it into a new sim that will be DX 10 compliant, multicore capable and will take advantage of ganged video cards.

[Paul J 349]

. . . oh - I forgot - you will able to smell the Jet-A burning when your favorite turbine lights up. . . . . .

[boshar 6]

All that doesn't realy matter much though. Joe Average will buy an off the shelves highend PC. Install the standard game. Maybe install the official patches (none for FS-X) and set the settings to high (thats no avarage PC they are testing there why shouldn't they activate all the eye candy).Thats whats tested here and FS-X performs horrible in that scenario. Other highend games fare beter.We may not like it but that doesn't mean it isn't true. FS-X is not well adapted to the mainstream market.Also note we are testing top of the range PC's here. Do you realy think that the middle of the range PC's on the shelves 2007 - 2008 will have similar or beter specs then that? They might have more processor cores but as we al know thats useless with FS-X.

[McCrash 3]

>http://www.extremetech.com/article2/0,1697,2059176,00.asp>>This is a pretty stunning indicator of how hard FSX actually>hits systems in terms of performance. >The tested systems run about $7,000, and given that FSX>delivered less one third of the performance of FEAR (which is>famous for delivering poor performance) on those machines and>never broke 20FPS, I'll be very interested to see what kind>of performance gain we get from patches.The flight simulator churns lots of numbers.Well the last and only patch for fs9 took a couple of years.

[Guest JackDanielsDrinker]

>FS brings nothing to its knees. FS wastes system cycles and>FS does *******************NOT************ use dual cores,>virtual cpus etc. in effect, FS is the one of the few>benchmarked games that runs on half the system. >>now that we've gotten this out of the way let's just relax and>move on... Really? So all the other games run on the "whole" system? They're all threaded to use multiple cores? Sorry, not true. They vary in how much they can use the second (or third...) core. Some of them might get 60-80% utilization out of the second core. Some get no utilization. FSX gets about 20-25% utilization. Don't look at framerates from one benchmark and make wild generalizations. FSX is completely different than those other games. First person shooters operate in a controlled environment in levels that are hand optimized by the developer. FSX has to be flexible enough to let you fly ANYWHERE ON EARTH. That is a huge performance challenge.First person shooters can control the view distance, where you can go, and what you can see so they can be optimized for higher frame rates. In FSX, you control where you go, and you don't need 60+ FPS cause you're not trying to twitch your way to a high body count.Also, first person shooters are not modeling weather, 3D or 2D panels, GPS, VOR, NDB, or ILS. They do model a person running around firing a gun. FSX models an airplane flying in three dimensions affected by wind, weather, gravity, drag, thrust, lift, your control inputs, etc. There is a lot of stuff going on here.In short, FSX has to do a lot more to make it "as real as it gets" than a first person shooter, where the scope of "as real as it gets" is limited to 10 or 20 hand crafted levels. Also, Extremetech probably turned up the Autogen to max which you shouldn't do. Maybe do this in a year or two, but not on today's processors. You're still going to see plenty of buildings when Autogen is set to normal and you will get OK frame rates on a good machine. You get to take advantage of tomorrow's processors when they arrive. Most first person shooters start to look dated after 2 years.