Archived

This topic is now archived and is closed to further replies.

Guest PowderBlue

Phils post about FPS

Recommended Posts

It never ceases to amaze me. I'll admit that I spoil myself for FSX and I go out and buy nice stuff but I'm still not happy. How is toms review getting 80 FPS when I've never seen 80 FPS regardless of what I try. Here are my new system specs 680 SLI LT from EVGA 2 X EVGA 8800 GTX O/C video cards in SLI 4 gigs Corsair XMS2 ram at 5-5-5-12 800mhz WD 150 gig 10000rpm HD SATA 300 gig HD Storage and to top it off running a Intel Core 2 Extreme QX6850 Kentsfield 3.0GHz 2 x 4MB L2 Cache LGA 775 130W Quad-Core Processor to run everything. Now on my computer with settings maxed and no AA I get about 12 FPS steady when my FPS slider is on 20 and I get about 20 to 22 FPS variably with slider set to unlimited. This is with running FEX, UTX, GEX but no activesky activated or any other 3rd party programs. Phil and I'm not knocking him but explains to us that toms review will make easier for us to understand that those types of numbers are obtainable but I can never seem to reach them even with this setup. One thing to me thats hard to understand is setting up the GPU in the Nvidia control panel. Do I leave it on single GPU with or multiple display? Cause I actually get better performance with it on multiple but I'm not running multiple monitors or windows. Do you leave it on single gpu or do you force split frames? It just seems irritating that after spending this money that you still have to tweak it out and do this and do that. Doesn't raw power just mean raw power. Any help would be much appreciated

Share this post


Link to post
Share on other sites
Help AVSIM continue to serve you!
Please donate today!

Hi,I set all my scenery sliders to max and ticked everey feature, DX9 mode. AI at 100%. One cloud (Stratus) layer only. Visibility at 40nm.At Copenhagen Kastrup Airport (EKCH) in the Default C172 VC I got an average fps of 13 standing still on the runway.I can't run FSX with smooth fps above 20 with these settings. So I gues you also have to move some sliders to the left and untick some check boxes.Did Tom's Hardware get 80 fps with all settings maxed out? I have a hard time believing that.Ulf BCore2Duo X6800 at 3.3GHz4GB RAM Corsair XMS2-8500C5BFG 8800GTX (OC:ed with RivaTuner), Creative SB X-FiFSX Acc/SP2, Vista 32

Share this post


Link to post
Share on other sites

It was on Ultra settings here is the review, I find it hard to believe myself, I'm not ssaying he is lying but come one If you have a clean install, latest drivers and all this equipment and your saying right out of the box you can't hold a steady 15 to 20 FPS anywhere how are you gonna expect anything morehttp://www.tomshardware.com/reviews/cpu-gp...de,1928-10.html

Share this post


Link to post
Share on other sites

GuysLOOK at the test parametersNO AA NO AFINgame AA/AFThey tested certain locations/missions, NOT free flightRegardless of what Toms Hardware is up to.. and they did 'forget' the 8800GTX 768 384bit cards... what Phil was pointing out is the relationship between the card and the proc. He was also pointing out out far things had come in just over a year... and then went on to touch briefly on the future and x58.His point was, FSX is coming into its hardware perf 'time' now and the importance of matching the card to the proc and visa-versaJust like FS9 took almost 2 years... FSX is delivering MORE than FS9 in less than 2 years and should start having the hardware needed to really make it sing in the next gen of releases at the end of this year... which puts it about the same as FS9 when you consider how long it took for hardware to catch up to the title.

Share this post


Link to post
Share on other sites

"How is Toms review getting 80 FPS when I've never seen 80 FPS regardless of what I try."You must not be trying very hard. I have a single-core Pentium 3.4Ghz with 2Gb RAM and an nVidia 7950. So, my rig is far inferior to your rig.Here's a screenshot that I just took showing 154.5 frames per second on the default flight:http://forums.avsim.net/user_files/188941.jpgKeep in mind, all I was trying to do here is maximize my frames per second without regard to anything else, which you said you tried to do.Here's Sitka Approach (the tested scenario) ... showing 80 frames per second.http://forums.avsim.net/user_files/188942.jpgIt's not very pretty, and I wouldn't want to fly with these settings all the time, but getting 80 FPS seems to me to be a trivial pursuit. I can fly with very high settings on and get 40-50 FPS on this mission. (Email me, I'll send you a settings .cfg file.)Of course, these very high FPS shots are with settings Max Low (not max high) but just showing what kind of FSP is possible here.Toms never explicity says they had all their sliders to the right. They use the term "Ultra Quality" which, I have no idea what that term is supposed to mean. It's meaningless in the taxonomy of FSX settings, and they don't say what it means, so I think trying to make comparisons from their test to your box is fraught with danger.They're making an apples-apples comparison between their boxes, so from that standpoint, it doesn't really matter what their settings are, as the only thing you can surmise from their test is the relative performance of the hardware they chose, and even at that, their testing methodology isn't scientific, nor available, and the article has its facts wrong about DX9c.So, I think you should be happy with your rig. I'd kill for it.

Share this post


Link to post
Share on other sites

I can get jumps to 80 fps with FSX SP2 Vista,default scenery, max clouds, max traffic, with737PIC for FSX, away from big cities. No autogen.some water effects2 Core Duo E6600 Gf 8800GTS 320mb 4 gig ramAA 4x AF 8x no tweaksPeter Sydney Australia

Share this post


Link to post
Share on other sites

Consider creeping up on the FPS issue slowly. FSX will run at max eyecandy with any of the default airplanes. For now, define 'eyecandy' as all the sliders on the left side of the scenery tab + all the driver based AA, AF, ETC engaged. To put this in a (Very) general context, consider that the following is All Vcard. To start, set the Vcard up properly per Nick's nHancer suggestion (AA to 8, AF to 16 and the rest). Then set all the sliders on the left side of the scenery tab to full blast (water a notch down, if you want). Then, completely shut down (Totally OFF) all autogen, special effects and AI traffic. Fly. There's your 80 FPS. On my Q @ 3.6/9800GTX, I see above 100 all the time.That setup is showing your Vcard's capability . . . and also shows that modern Vcards have No Trouble handling anything FSX can throw at it. Any of the 88-98 cards are Plenty for FSX.Now start bringing in the CPU. Start increasing special effects, AG and AI levels. Your frame rates will tank in a blink. THIS is where the bottleneck is. We need more processing capability, be it optimized memory subsystems, more intelligent multi-CPU engagement or just brute CPU clock. At this point, a $150 8800GT will provide 100% of the function that any Vcard can provide for FSX (resolution dependent). Strictly relative to FSX generated loads, more Vcard will not help. No mix and match or intuitive analysis required. Get an 8800GT and wait for the processing bottleneck to resolve. That's the best anyone can do right now. I'm not optimistic about a large performance increase occurring by the means of Any next-gen hardware improvement. Nick's demos with next-gen capabilities looked promising, but can these be brought on-line so we-the-people can just plug and play? Even if they can, these will be important, but Only incremental improvements. Good thing the PMDG/FS9 setup is mostly in the cockpit. It's not very pretty out there even with FS9's megascenery. We're spoiled (or at least I am). The big boost will come from optimized, multi-thread aware software . . . and it can't come soon enough. I'm certain this at least one of the primary goals of the FS11 dev crew.

Share this post


Link to post
Share on other sites

I was not estimating FSX performance but I was estimating real world perf based on what I have seen coming out of Asia and a few other places. We all know FSX is very CPU bound and anything that uncorks the bottlenecks with the current proc/chipset design is going to improve the FSX experience.As for what to upgrade to between now and then... well, that has to do with what you are using now and how you are using it. If you already have a quad core based on C2 technology, upgrading to a 'tick' or 45nm proc is not going to get you much, especially if you do not clock. At the same time in this day and age with Intel, overclocking is practically a prerequisite to a better FSX. Intel placed anywhere from 500- 1000 dollars worth of overhead into their cores so not tapping into that is costing. If you do clock and already have a good quad, the money should probably go into a card.. That choice would also depend on what you are currently running. Personally, I would not buy the current offerings due to cost/performance ratio as compared to what I run now however that could change this summer. If I was not running a 768 384bit card I would probably look at the upgrade sooner however as it stands.. anything I throw money at right now is now going to provide much more than a marginal change therefore I would wait a bit longer and look at what hits the streets this summer.Anyone who may be coming from less that what I posted above probably should take a good look at the chart Phil pointed to. Mating the right proc/card will make a difference.

Share this post


Link to post
Share on other sites

Try the Sitka Appr Mission looks like that is what TH used.Just load it, pause, Shift Z to bring up framerate, then release note your fps and setup...With a 4Ghz processor and a 8800GTS 640 I'm seeing 43-48 with very high quality settings at 1920x1200 in game AA/AF DX10.Tom's reporting 71 ! with a 8800GT at 2.94Ghz????Seems impossible with any settings that would offer decent image quality, what do you get?

Share this post


Link to post
Share on other sites

I think you might have to have the DX10 check box checked to see the hight frames. Yesterday I was at a solid 60fps over default fsx with medium autogen enabled, and my computer is way less powerful then yours.

Share this post


Link to post
Share on other sites

>>> It never ceases to amaze me. I'll admit that I spoil myself>for FSX and I go out and buy nice stuff but I'm still not>happy.>> How is toms review getting 80 FPS when I've never seen 80 FPS>regardless of what I try.>> Here are my new system specs >> 680 SLI LT from EVGA> 2 X EVGA 8800 GTX O/C video cards in SLI> 4 gigs Corsair XMS2 ram at 5-5-5-12 800mhz> WD 150 gig 10000rpm HD> SATA 300 gig HD Storage> and to top it off running a Intel Core 2 Extreme QX6850>Kentsfield 3.0GHz 2 x 4MB L2 Cache LGA 775 130W Quad-Core>Processor to run everything. >> Now on my computer with settings maxed and no AA I get about>12 FPS steady when my FPS slider is on 20 and I get about 20>to 22 FPS variably with slider set to unlimited. This is with>running FEX, UTX, GEX but no activesky activated or any other>3rd party programs.>>> Phil and I'm not knocking him but explains to us that toms>review will make easier for us to understand that those types>of numbers are obtainable but I can never seem to reach them>even with this setup.>> One thing to me thats hard to understand is setting up the>GPU in the Nvidia control panel. Do I leave it on single GPU>with or multiple display? Cause I actually get better>performance with it on multiple but I'm not running multiple>monitors or windows. Do you leave it on single gpu or do you>force split frames? It just seems irritating that after>spending this money that you still have to tweak it out and do>this and do that. Doesn't raw power just mean raw power.>> Any help would be much appreciated>If your comfortable flashing your mobo flash it with the latest 680i BIOS (not 680i LT), then overclock that CPU and enjoy +FPS.

Share this post


Link to post
Share on other sites

I'm not seeing the 71 FPS with mine either. I'm getting a throughly jaggy ~ 22 FPS without AA (UgLy) and a 2 FPS drop to 20 with 8xS engaged. For me, AA is virtually free. Notice Tom's chart simply shows that FPS tracks directly with CPU speed. The (8800/9800) Vcard/CPU match-ups are virtually meaningless. FPS ranges are all CPU driven. Nothing new there. Just get a $150 8800GT and ride it out. http://www.tomshardware.com/reviews/cpu-gp...de,1928-10.htmlBut where is that ~ 80 FPS coming from? I agree that something's going on. Frankly, I don't believe any system at less than 5Ghz! can get 70FPS with all tab presets to Ultra presets. How did they get it. My bet is that they may not have had All the tabs pre-set to Ultra. On their VGA comparison page for FSX at 19x10, Tom's reports 22FPS. http://www.tomshardware.com/charts/That's just what I get too at the same settings (Q @ 3.6 + an nHanced 9800GTX @ 1920/1080).For better comparisons, I maintained altitude but turned 180 and flew back into consistent terrain (Frame rates were going off the chart as I headed out to sea).Traffic AI to Off gains nothing. Guess there's not much traffic out there. Eastbound over consistent terrain: Weather to Off gained 10FPS to ~ 30FPS.On the Scenery tab, All "Scenery Effects" to off gained 40FPS to ~ 70FPS. -- Right here. Setting all "Scenery Effects" to Off unloads the CPU bottleneck and frame rates really begin to roll. Tom may have left this setting at "off."On the Scenery tab, All "Terrain and water" to off gained another 80FPS to ~ 150FPS.The aircraft tab has no effect. What cha think?

Share this post


Link to post
Share on other sites

Based on my specs, I thought I should be getting over 70fps on the Sitka mission with all settings at "Ultra" high under DX10. But I could only get about 25. I'm glad I read this thread before spending hours trying to figure out what was uniquely wrong with my system.It turns out that the Tom's review is seriously wrong and/or misleading. And it's no excuse that "they were just trying to show relativities": unless you can trust the reported framerates, you've no reason to trust the relativities, either.Tim(Relevant specs: 2 x Xeon 5160 (equivalent to 2 x 3GHz Core 2 Duo), 8800 Ultra "factory super-overclocked" by EVGA, 4Gb RAM, 15k SCSI RAID0)

Share this post


Link to post
Share on other sites

I think the real takeaway is that any modern 88/98 Vcard is fine. Any 88/98 will provide 100% of Any performance gain that will be available by means of Any Vcard. If you want more FPS right now, upclock a Single core. Tom's goes on and on about the meaningful difference between 70 and 74 FPS. This is ridiculous. We need to recognize very clearly "That Nvidian King had no Cloths!" It's the same ol' rap. Watch out for these salespeople. I'm very interested to see what difference the new GTX280 will make. We'll just have to see. Only a couple of weeks to go.http://forums.vr-zone.com/showthread.php?t=280553

Share this post


Link to post
Share on other sites

"Frankly, I don't believe any system at less than 5Ghz! can get 70FPS with all tab presets to Ultra presets. How did they get it. My bet is that they may not have had All the tabs pre-set to Ultra."They never claimed that was their test setup.All they wrote is this:* 0xAA (let's assume this is anti-aliasing is off)* Trilinear filtering* Ultra QualityWhat does "Ultra Quality" mean? What "ultra quality" means is anyone's guess, but clearly it does not mean "every slider to the right." I would bet, based on the quality of the reporting vis a vis DX9c, that all it means is that the tester pushed one slider to the right ... just this one:http://forums.avsim.net/user_files/188967.jpgIf that's true, they didn't push all sliders to the right. They only pushed one slider to the right. But, we'll never know, since it would be trivial for them to release their test settings.cfg file, but they don't.Nevertheless, assuming they applied the same test to all the machines (and that's a huge assumption), then all you can take away from this article is the relative performance of the devices to each other, and you have to ignore the raw FPS numbers.They're meaningless. As Phil pointed out when he said: "Its a benchmark of systems against each other. As such, showing the high-water mark shows where the variants stand against each other and that has value for people considering a purchase. As to the 'value' of any particular result, that isn't really the point. It's the whole, not the parts."I'm guessing that was Phil's way of saying: "Ignore the 80 FPS. Instead, compare the systems against each other to decide which is 'better' and how far they've come."On that basis, the article has merit (assuming, of course, they did the test correctly.) What it clearly does not mean is that you will get 80 FPS using their machine flying with your settings configuration.Commissioner Gordon out.

Share this post


Link to post
Share on other sites

The way I understand it Phil pointed the article out not because of the specific framerates being achieved but to illustrate the improvement in framerates using the newer hardware.

Share this post


Link to post
Share on other sites

In Toms review, wheres the 8800GTX and 8800 Ultra?The wider memory bus should have put them on top of the FSX list.

Share this post


Link to post
Share on other sites

Sorry, but I think the Tom's article is completely worthless so far as insight into FSX is concerned. As the posts on this thread show, nobody can now have any confidence that the same settings were used with different setups. The clear implication of the article is that all settings were at "Ultra" high: but that is manifest nonsense, as anyone with a modern PC who has tried the Sitka approach at "Ultra" high settings will confirm in an instant.Since the article is unreliable in these crucial respects, it seems to me that it is unreliable in all respects. That is the trouble with careless journalism: once it becomes clear that PART of an article is lacking in rigour, it becomes difficult to place confidence in ANY of it. Makes you wonder what was the purpose of the article.By the way, I'm not having a go at poor old Phil Taylor, who obviously feels under the kosh at the moment: the article was nothing to do with him.Tim

Share this post


Link to post
Share on other sites

I have been able to attain 209.2 FPS in the same area and same FS setup as Tom's Hardware. I set an average of almost 21,000 3dmarks vs 15,000 in the THW highest end spec system. I suggest you start there.See my specs below.Regards,Mike T.

Share this post


Link to post
Share on other sites

That's just a Q at 4. Xfire/SLI bumps eyecandy, but won't help with FPS. Check your sliders.

Share this post


Link to post
Share on other sites