Benchmarking FS2002 for Frame Rates- Anybody doing this

[brucek 38]

I've recently built a new system and before I start making too many changes would like to benchmark it. I have SiSoft Sandra, although I'm not sure what most of the benchmarks mean yet. :)What better though than what really matters: FS2002 performance. I'm thinking that to benchmark FS., you'd need to first remove any cap on fps that has been set up on Options/Display, and maybe even disable AI traffic, as it varies not only by day and hour, but there's so much development in this area that traffic.bgl's are liable to big changes. (Note, my thoughts are for my internal use in tweaking alone, although I'd be more than happy to share with anyone).Any ideas, or is anyone doing this already?Thanks,Bruce.

[barryward12345 0]

Somebody "tried" to do this when FS2002 first came out -- but there was so much disaggreement and "noise" generated in the forum that it seemed to die.Personally , I believe that the only way it could be done would be to adopt these settings 1 Default Meigs startup with default Cessna 172 (everybody has this)2 No Weather (hence no clouds, etc) 3 No ATC (for the reasons stated by you) 4 All settings set to MAX (regardless of whether you normally fly like this or not!)5 Framerates set to unlimited Then see what you get.-- after I disconnect from the Internet , I will do this myself and report back. Trouble is a that a permanent home is needed for any results reported. I could actually do this - setup a webpage -- but in a few weeks time , I will be off-air for 6 weeks.Barry

[Guest]

Bruce,Your methodology is probably the only accurate way of doing any form of benchmarking within FS2002, as any frame rate settings cap will probably result in sliders being moved behind the scenes.Additionally, to make the test complete, you need to probably disable not only the dynamic ATC but also the dynamic scenery and weather as it can introduce randomness into the testing. Add to this a finite flight plan over the same terrain coupled with test measure points at specific locations (ie: dme 30, 25, 20, 15, 10, 5 when flying towards a VOR) and you can claim what should be a fairly consistant testbed.The problem is that all this dynamic content is what over the years has made FS grow and prosper, without it, things are somewhat boring, as we want company, nasty weather with huge scary clouds, the odd windshear and gusting winds :-lolAny other benchmarking outside of FS2002 really means nothing at all, even things like 3DMark can be made to vary a few hundred to a few thousand in results due to a minor feature change in a tweeking tool. These external benchmark tools are handy at looking into why your machine is not measuring up compared to others with the same hardware, but even that has to be taken with a grain of salt as people overclock their hardware far beyond the base performance that it should provide in native configuration and post these results which lead the normal folks into believing that their system is not working properly.Ray

[brucek 38]

Thanks Ray and Baz. Some good ideas. I hope to try these tonight or tomorrow.Bruce.

[barryward12345 0]

>Thanks Ray and Baz. Some good ideas. I hope to try these >tonight or tomorrow. >>Bruce. Reporting back as promised .With the conditions set as per my previous post I am getting 12.5I have a P3-1000 512 meg SD RAM and Geforce 2MX video.When I cut back "slightly" on some of the settings , I get 16 which is what I lock it to when flying.Barry

[Guest]

Hello Bruce,Actually there was an FS benchmarksite, and a very good one too, right here at AVSIM. It still exists: go to "Partners" and choose "FS Bench". It was set up a long time ago by Bruce Wilson, later managed by Pierre Poirier. I was involved a little as well. My last contribution being a final analysis of the FS2000 benchmarks. Conclusion: FS2000, like its predecessors, is computebound. With respect to performance (framerate) the only thing that really matters (linearly) is cpu-power! :D However videocards, memory etc. can add to the picture quality and overall smoothness.It was intended that there would be a new benchmark for FS2002 as well, but unfortunately that never happened. Some of the reasons are mentioned in this thread. Benchmarks only make sense (are only valid) in a comparable and stable environment. The problem being that there are so many variables in FS2002, that it's very difficult to create such a stable environment. The other problem being lack of time by Pierre :-( I don't know of any other benchmark available. Bubba Wolford at Simulation Community Headquarters (http://www.simhq.com/) has created one for himself to compare graphic cards, but I don't know whether a description is available and useful for your (our?) purpose. BTW, he too concludes that (for FS2002) graphics cards only add to the image quality and smoothness, but not much to performance! Not even a GeForce 4 Ti 4600 !:~P If anybody feels like filling the gap, let him please contact Pierre Poirier at "FS Bench". I am afraid this doesn't help you much, however,Good luck, Jos G.FS History (http://simflight.com/fshistory)

[brucek 38]

Interesting result. ( I spent all of last night installing an OS onto my wife's machine, so am still to get to this, but am very interested).Thanks,Bruce.

[brucek 38]

Thanks for the info, Jos. Yes, I recall Bruce (great name!) Wilson, it was about the time when AGP and Win95 OSR2 came out, if I recall correctly.Thanks again, I'll write Bubba.Bruce.

[brynjames 0]

I use the FS2000 benchmark tests which are still here on avsim at http://avsim.com/fsbench/fsb-fs2k.htm They may not be perfect for FS2002 but they have enabled me to see the improvements (or otherwise) when changing components in my systems.As you say, turn off AI, but otherwise take the weather and other setting suggested by the FSBench authors.

[Francois_Dumas 10]

Hi Bruce, Jos, et all,Yes Pierre did write me a few times. He lost the access to that database here at AVSIM and so far we have not seem to be able to get him that access again.I beleive he has given up now :-( Don't know what happened there, really.I'd surely be interested in such kind of benchmarking too and would be happy to try and see if we could host a new 'site' here at AVSIM with all results. I am sure it would be beneficial in the recurring discussions about this topic ;-)Regards,Francois________________________Francois A. "Navman" DumasAssociate Editor &Forums AdministratorAVSIM Online!email: fdumas@avsim.com________________________

[Guest B1900 Mech]

20.3 FPS On a Athlon 1.3,512Mg PC2100 DDR, Elsa GF3 64 Mg, 7200 rpm Hd, Win-Me, 1024-768-32.

[barryward12345 0]

>20.3 FPS On a Athlon 1.3,512Mg PC2100 DDR, Elsa GF3 64 Mg, >7200 rpm Hd, Win-Me, 1024-768-32. Now - this is good .Was this done in accordance with what I suggested in post 2?If we can get some more reports like this then they can be collated.You get 20.3 - I get 12.5. The big problem I guess is that we don't know just what bits of your machine are giving you this better result. Obviously your Athlon 1300 is faster than my Pentium 1000 -- but I wonder what contribution your PC2100 ram gives over my PC133 SDRAM. Obviously some - but how can we quantify it?Barry

[Guest B1900 Mech]

Hi Barry! As I am no computer expert I can't say, I wonder if the screen resolution should be one standard? P.S. I folowed your standard.

[Guest B1900 Mech]

Bump, What about cockpit veiw setting, for example .70, also a standard vidio setting.

[Guest Stamatis]

Barry,You say "all settings to MAX". OK, this covers the options set via sliders. What about the ON/OFF menu options, you know, the little square boxes where you are suppossed to tick them ON or leave them empty (OFF). Do you also check all these options ON?Stamatis