I'm doing a lot of benchmarks these days since I've got a new computer and fresh and uncluttered installs of FSX and P3D. This weekend I've been doing extensive tests using my standard 15 minutes flight crisscrossing New York from KJFK to KEWR. The i7 2700K has a default turbo mode of 3,9GHz. During testing I've pushed it to 4,2 and 4,6GHz. The computer is very stable at these speeds as well. Surprisingly though I've found overclocking to have very little impact on performance. Average fps increased only 2 fps from 3,9 to 4,6 GHz (18,5 to 20,7) to be precise. I'm posting the full benchmark on my blog later this week so you can see my settings.Anyway, I thought the results were interesting since I've always heard that overclocking is a "must" for a smooth ride in FSX. Of course, fps doesn't really tell how well the sim runs, but it is the only objective unit of measure we have. Subjectively I didn't notice any difference in terms of stutters and smoothness when the computer was overclocked. I'm nowgoing to do a test using the affinitymask tweak and disabele hyperthreading to see how that impacts performance.

Edited February 27, 201214 yr by simmerhead

I guess it can only add to Stephen's body of work.https://docs.google.com/spreadsheet/ccc?authkey=CMSSocoI&key=0AuiXjW8fJyDrdC1KZm5fSlJYdnJBaXRRclFuU0U3WUE&hl=en&authkey=CMSSocoI#gid=0 and:- http://www.lockonfiles.com/index.php/topic/33485-480-gtx-vs-580-gtx-review/

Well, I don't know about your experience but the difference between the 2500k @ 3.4 and 2500k @5 ghz is unbeleviable.

Thanks for the links Paul! Stephen has done some interesting testing for sure. I haven't gone the FSXMark11 route since I want to test FSX on my system and with my preferred settings. Don't think there is much reason to add to that spreadsheet. I use the same methodology though using FRAPS to measure an autopiloted flight. Ron, how do you notice the difference?

fresh and uncluttered installs of FSX and P3D

The clue is in the statement "fresh and uncluttered installs of FSX and P3D"Throw all of your addons into the mix and report back with your findings

The clue is in the statement "fresh and uncluttered installs of FSX and P3D"Throw all of your addons into the mix and report back with your findings

Indeed. That is why I am building up a library of benchmarks. The problem of corse is that it is almost impossible to measure how addons affect eacother.

..and it'll still be impossible, SimmerHead. Everyone else's power supply is different: GPU's can be plugged into different slots: monitors are different: mobos are different: the permutations are enormous. 30 MB's x 30 PSU's x 15 connectors x 10 580's x 4 slots, x 40 monitors x all the different GPU, Mouse, USB and Audio driver combinations? Disks? Did I mention disks!!You are carrying out a futile task, my friend: we already know how much of a difference a few parts of a gig impact ones ability to move sliders around. The sweet spot on a SB paired with a 580 is 'around' 4.5 - 4.6 gig: above that - within some minor limitation - one can configure FSX as one likes. There are a good number of guys here that are fortunate enough to be in that spot, and we all agree with these observations. Ron Hamilton's is perfectly correct.

Very interesting.... When I sent my chip to 4.8GHZ, the performance boost was massive.I saw about a 10-15FPS jump in certain scenarios....

Overclocking my 2500k was also a huge improvement. Not really in overall FPS but in what addons I could.. well.. add on, before performance tanked.-G

I noticed the largest increase running more AA and AF while using high LOD, like 6.5 in Orbx land... With default scenery/planes you probably won't notice a huge difference between 3.9 and 4.6. Start adding fps hogging payware planes and scenery, traffic, high levels of driver-level video card stuff and you'll notice the extra MHz helping you out.

Start adding fps hogging payware

The NGX really seemed to take advantage of the higher clock speed. I saw a very large jump in performance in the NGX from default to [email protected].

Surprisingly though I've found overclocking to have very little impact on performance. Average fps increased only 2 fps from 3,9 to 4,6 GHz (18,5 to 20,7) to be precise.

That sure sounds like you experienced a different bottleneck in your system: PCIe, the GPU, both, etc . . .

That sure sounds like you experienced a different bottleneck in your system: PCIe, the GPU, both, etc . . .

That might be it.It is futile indeed Paul, but as I've learned one has to test FSX on ones own computer since there are so many different findings out there as to how FSX should be optimized. From what others have experienced it will be interesting to throw addons into the mix. Hopefully overclocking will decrease the fps drop I would normally get?

Max I can seem to get to is 4.6, any notch higher and I get warnings straight from bootup. But thats mainly because I don't have the greatest mobo, a gigabyte z68. If I knew 6 months ago I would have got the P68 Extreme and I'd really be able to use the i7 to its full potential.

My 2700K works for FSX almost identical between stock and 4.8GHz. When windowed, small screen I can see advantage from 190-210FPS to 200-220. That's it. I was in Barcelona's El Prat yesterday - full screen results: 17 FPS unclocked, 18-19 with 4.8GHz. Slight difference, but worth IMO.