After the unfortunate demise of the Petraeus Index from the avsim hack last year I decided to re-create my results with more recent testing. While this isn't nearly as good as the Index was (because they had multiple participants and took the average result) it is decent and hopefully what you might expect if you were to purchase and install an aircraft.What this chart shows is a percentage of FPS against a baseline test. The baseline test is the default 737 VC, set to parameters per picture below.The reason I am fascinated with the Petraeus Index is because this data leaves out major factors like hardware. A lot of people say "Oh I get 40 fps with XYZ aircraft!" This doesn't mean jack to someone running a completely different system! This way, one tests all their addons against the default 737 so they can decipher performance locally.Te theory behind the original index was to have multiple participants, then calculate the average, then display a chart.Obviously I only have my results here but I'll upload the Excel files if you want or you can create your own.I realized after doing this that I've forgotten about 10 other payware and freeware addons. They are not in my "hangar" at the moment so I should get to them later on.One thing I added is FPS variance. In my testing, I've noticed planes with higher variance have produced more stutters, especially while panning in the cockpit and turning. So I added that as well...DATA + CHART:

Which Citation X? v1.0 or v2.0?Also... system specifications?You're leaving out important information. :unsure:Things I know affect frame rates:1 - Scenery complexity. For testing all scenery settings should be at absolute minimum.2 - AI. For testing all traffic settings should be at absolute minimum.3 - Drive access. For testing accuracy non FPS measurements should be made until FSX has finished loading data.4 - CPU/Video card. These two items can have significant impact in performance results.5 - Background tasks. These can also impact frame rates significantly.

HelloI get way better frames on the Eaglesoft CX v2 compared to the F1 Mustang to the point where the Mustang is not a pleasure to use on my hardware.

Which Citation X? v1.0 or v2.0?Also... system specifications?

2.0 - and the system specs don't matter in this test. The reason is because I'm testing everything against the default planes - on my own system. Which is exactly what other users would do...Once you have plenty of people testing on their own systems it will create an average. The point is not to say "I have 30 fps in the CitationX , while Fred has 25 fps." It is to say I see a 55% drop in fps on MY system compared to default 737. Someone else might have only an 35% drop because they're on a Core i7 yes but once you average everything out (with multiple participants) there will be a FPS drop on average between different products.

HelloI get way better frames on the Eaglesoft CX v2 compared to the F1 Mustang to the point where the Mustang is not a pleasure to use on my hardware.

Can you add to the test? I'd like to add as many results as possible...File avail here:http://cid-408ca931940d5e3c.office.live.com/view.aspx/.Public/PetraeusIndexAug2010.xls

The index is flawed. It isn't being tested to represent actual performance. All the steps I listed in my post are a must if you want to be able to define the aircraft's performance.Without elimination of the other items, performance values will never accurately reflect the aircraft's performance at all.

The purpose of the orinal Petraeus index as ryanbatcund pointed out was away to compare all aircraft to the performance of that of a "baseline" aircraft. ryanbatcund has choosen to use the stock 737 performance to compare all other aircraft to. The index was to designed to show a percent difference in performance from the "baseline" to an aircraft being compared to it. The idea being that no matter what hardware or addons you may be using the percent difference in performance between the "baseline" aircraft and the compared aircraft would remain consistent. FPS means nothing in this index, and I beleive ryanbatcund just happend to include that for reference. System specs, addons, hardware etc. will all affect frame rate but when comparing the performance of the "baseline" aircraft to the compared aircraft they mean nothing and what was found with the original index is that overall aircraft performance (percent difference from baseline) will remain fairly consistent regardless of system specs, hardware or addons. In other words FPS may vary with system spec, addons and hardware but the percent difference in performance of any given aircraft to the "baseline" aircraft will remain consistent. I'm glad to see ryanbatcund resurecting the idea. It was a very useful tool. I apologize for the above being so wordy.Regards,Bobryanbatcund,Have you tried to contact Petraeus to see if he has any of the orginal data?

The index is flawed. It isn't being tested to represent actual performance. All the steps I listed in my post are a must if you want to be able to define the aircraft's performance.Without elimination of the other items, performance values will never accurately reflect the aircraft's performance at all.

Of course it won't be perfect but the test is a broad scope of performance percentages.The point is not to obtain actual FPS, it's to see a trend or ratio of performance against a plane that everyone has in their inventory. If you test with the same procedures everytime you'll get roughly the same result everytime. So if you make sure all your hardware is neat and tidy, and your AV is off etc etc, you'll get the same results on your system everytime... (within reason, there's always going to be error)

FPS means nothing in this index, and I beleive ryanbatcund just happend to include that for reference. System specs, addons, hardware etc. will all affect frame rate but when comparing the performance of the "baseline" aircraft to the compared aircraft they mean nothing and what was found with the original index is that overall aircraft performance (percent difference from basline) will remain fairly consistent regardless of system specs, hardware or addons.

Yes.... perhaps I should have used 100 as the baseline like they did in the original one. FPS don't mean good or bad, they were just a number...less calculations this way :(

The default aircraft will perform differently (drastically so) based on the items I've listed in my earlier reply.If you refuse to eliminate those items, you will never be able to represent how the actual aircraft's physical performance is.As example.. the 3D world load is primarily based upon a user's hardware: memory, CPU, video card/memory, drive access times, drive clarity (defragged, etc).None of those is being eliminated in the test, thus the test values are completely invalid as a representation of how an aircraft addon performs on a user's system. What you're really showing with this test is how well a system performs with everything in FSX active, including a given aircraft addon. It doesn't represent the aircraft's performance whatsoever.

Ed,I'm not going to disagree with you in the least. However, as I said when this was done before the percent difference to a "baseline" among systems was found to be fairly consistent. There were also some constraints on where the testing was to be performed (sitting on a runway, at a particular airfield), how it was to be performed (idling), rebooting your computer and FSX prior to each run etc. It was a good tool to get an idea one aircraft's performance relative to another.

It doesn't represent the aircraft's performance whatsoever.

It does relative to your system, and other aircraft you have in your inventory.For instance, the RealAir Duke will perform similar to the Eaglesoft SR22, across a wide variety of systems. The C750 and F1 Mustang will display lower fps than most GA aircraft... etc etc....The Carenado aircraft will be on par with default performance, if not better...Doesn't matter if you're running a Core i7 clocked to 4.0GHz or a Pentium 4 at 3.0Ghz - you will notice a FPS loss compared to the baseline B737. The chart will let you know what to reasonably expect, percentage-wise. It looks bad since it's only one user. I saw the original tests and graphs and they were a fairly accurate indicator of performance from about 10 different users' perspectives....

Just an observation of this performance index. Its only really useful when there's a very large amount of participants sending in data. Imagine this example.An addon aircraft is using 700 megs of GFX RAM, with an old system with only 500 megs of GFX ram you would get alot of swapping between GFX ram and system ram / swap space. This would drastically decrease the performance of this aircraft. The default 737 would not be affected by this gfx memory bottleneck, as its not using anywhere near the same amount of GFX memory. Now, say you have a 1gb gfx memory card, there would be no need to swap those textures in and out of memory and the performance would be massively improved compared to the 500 meg system. The percentage difference between the default 737 and the high gfx memory usage plane would be drastically different between these two systems. That is why the performance index needs a very large sample size, and even then the number would be unlikely to represent how it would perform on a real system. The performance index would either indicate a bigger difference (if you have a 1gig card) or a smaller difference (if you have a 500 meg card). It would never indicate the actual performance difference in such a scenario

It does relative to your system, and other aircraft you have in your inventory.

And there's where things go 'wrong'. I know I can change an aircraft's performance on my system. I can get different frame rates for the same, identical aircraft... because I know exactly where the largest framerate impacts come from. By changing system configuration and/or FSX settings I can make an aircraft appear to be a hot performer or a complete dog. Thus the 'index' you seek to define is completely invalid. The items that impact performance the most aren't even being eliminated in the testing.

I can get different frame rates for the same, identical aircraft...

Well I can't. Setup the flight same way, everytime. Sure you might get 5% error but for me it has been consistent.I don't think the average simmer knows how to get different fps for the same situation either...If these charts had no validity, then one could argue you'll get the same performance between the default 737, Carenado Arrow, and Eaglesoft C750 under the same testing procedures. No way they'll be in the same performance band!

Great post! Thanks , Ryan.

Well I can't. Setup the flight same way, everytime. Sure you might get 5% error but for me it has been consistent.I don't think the average simmer knows how to get different fps for the same situation either...If these charts had no validity, then one could argue you'll get the same performance between the default 737, Carenado Arrow, and Eaglesoft C750 under the same testing procedures. No way they'll be in the same performance band!

None of you are running 'the same situation'. That's the problem. 3D terrain and building rendering is the largest performance impact in FS9... it's also a big hit in FSX, but AI in FSX is a much larget performance impact. None of this have anything to do with the aircraft being tested, but have absolutely everything to do with hardware and software settings/capabilities.