Gary! Good info, thanks for your efforts here! If you are making a general benchmark for everyone, I think the best way is to use a recorded flight, rather than a saved flight like the original one posted in the Hardware forum was. With supplied benchmark settings file, this would insure no variables caused by movement would effect the outcome of one's benchmark. I think the best tool for this, would be fsrecorder. It's free so everyone should be able to use it, and it runs more stable than the FSX default recorder.

It was mostly in jest Phil. I'm glad that you have an interest in this thread and are participating.Edit: It would be great if you could convince Intel to participate also.Thanks, Ted

Thanks a lot, Gary!It would be interesting for me, how a higher FSB would affect your results. I think the Penryn-based Quads with FSB 1333 should be the best choice for FSX or am i wrong? Or do we need a Intel Skulltrail system to get rid of all stutterers?^^NilsPS: Can somebody describe the difference between the Xeon and the Core2 CPUs in general? What is the difference between e.g. the Q6700 and the Xenon 3230 (besides the price)?

Great topic Gary, and very interesting stuff to read about. I also have a Q6600 @ 2.4GHZ. I'm wondering how high I'd be able to overclock this cpu. What program do you use to overclock your cpu. There are some overclocking programs that come with my mobo, but I can't really get much more out of it. What do you suggest?Eric

>Thanks a lot, Gary!>It would be interesting for me, how a higher FSB would affect>your results. >I think the Penryn-based Quads with FSB 1333 should be the>best choice for FSX or am i wrong? >Or do we need a Intel Skulltrail system to get rid of all>stutterers?^^>>Nils>>PS: >Can somebody describe the difference between the Xeon and the>Core2 CPUs in general? What is the difference between e.g. the>Q6700 and the Xenon 3230 (besides the price)?>You don't have to wait for a Penryn for a 1333FSB, most OC'd systems are already clocked at least that high, and many higher. For stock CPU's, the E6750,E6850 (The one I have) have 1333 FSB's as well as the quad core QX6850. The benefit of the Penryn should be it can clock at faster clock speeds as it should in theory be a cooler chip than a 65ns CPU like the C2d's

Gary,More interesting results. I wish I had some suggestions for you on the DX10 drop in performance, but I'm still running XP on a single core. One thing I was wondering is how effectively your system is using your 3 gigs of your RAM. Is that 4 sticks? Just a thought.It is nice to see that SP2 offers a performance improvement over SP1 in DX9. This is with Vista 64, so I don't know if I can assume that the same results will occur with XP.More excellent work, I am looking forward to seeing your research develop.Ted

One of the reasons I removed Acceleration from my system was the lower performance in DX10 , although the primary reason was texture compatibility issues with my older FS9 aircraft. Even though performance was still acceptable in most areas on my new system, I can see this being a problem on lower-to mid range systems. I've since went to a dual boot Vista/XP, so I'm thinking of putting back Acceleration on my Vista system to use Acceleration features such as the FA-18 and aircraft carrier operations. The one area SP2 did deliver on was a marked increase in performance with Bloom on. Now it only eats a couple of FPS instead of Half!!

Ted, >I just noticed the dual boot in your sig. If you are still>taking suggestions and you run out of things to test, it would>be interesting to see the difference between DX9 and DX10 in>Vista as well as the difference between Vista and XP in DX9.The problem with my XP FSX installation is that it is not stock, (in particular Ulitmate Traffic X, FS Genesis mesh / landclass and X Graphics change the default run quite significantly) so I am going to have to wipe it and reinstall to make a valid comparion. I will do that, but I XP is my safety net for my Vista 64 install. Having said that, apart from the unexplained FSX DX10 performance drop, Vista 64 is going along just swimmingly for me and I actually perfer to use it over XP now :-eek :-hah. So, I might be happy to mess up my XP installation in very short order. Stay tuned for the XP side of the benchmark runs.Gary

Thanks for your reply Tom,i already have a E6750 @3,2 GHz running with FSB 1600.My FPS are good, but i want a Quad to ged rid of all stutterers.Thats why i'm interested in how the FSB is affecting the overall performance, because i don't want "downgrading" with a Quad.Nils

The very first version I made of the benchmark was with the demo and was a recorded flight, albeit recorded using FSX itself. Play back was very stuttery, although the results were consistent, so I changed to an autopilot flight. I am aware that FSrecorder does a better job a providing smooth playback, but I having sunk a lot of time into the autopilot version and it is actually extremely repeatable across multiple systems I have tried it on. The key was to ensure the aircraft was in stable autopilot flight (ie. at desired alt, speed and heading) before starting the recording. As the weather file will be included in the benchmark pack I put together, variability across systems should be absolutely minimal.Gary

Nils,In my experience, FSB makes little difference to resultant FPS at a given clock speed, but clock speed itself makes a huge difference. I will try two runs at the same clock speed but different FSB to confirm this.GaryPS: when I run my Q6600 @ 3.6GHz the FSB is 1600MHz, which is way above the current 1333MHz max standard.

Ted,I run 2x1GB sticks and 2x512MB sticks. All 3GB shows up and is used by Vista 64, and when I load up FSX between runs, it only take about 3 seconds from icon to free flight screen and about 8 secs to loading the flight with no hard drive access, so FSX is caching quite nicely and Vista appears to be using all the memory.Gary

One thing I did notice with the DX10 versus DX9 runs is that DX10 water and lighting looks SO much better than DX9, even at global high default settings (ie. bloom off and water at 2.x mid). In fact, I find it extremely difficult to tell whether bloom is turned off in DX10, as it looks like it is on all the time and performance doesn't seem to change between it being on and off (haven't benchmarked this - yet!). Perhaps bloom is on all the time in DX10 and that may explain why DX10 performance is lower?Gary

Eric, The quick answer is that I use the BIOS to overclock up to 3.5GHz and EasyTune5 to go the last 100MHz (because my computer won't boot at 3.6GHz, but it certainly runs stable at that speed once up and running). If you want to discuss OCing further, just head on over to the hardware forum and read what I have already posted there on the topic and ask any further questions if you have them. Gary

Gary,Great job! Thanks for taking the time to do this.:-)I wonder though, why alot of people are using 3GB RAM as opposed to 4GB, even though FS my not use it? Surely it would use at least what it needs?I also can't wait to see how my new QX6850 (8MB,3.67GHz Factory overclocked) does with 4GB Corsair Dominator DDR2 SDRAM 800MHz OC'd to 1066MHz-4 DIMMs on Vista Ultimate 32, with a Dual 768MB Nvidia GeForce 8800 Ultra in the new Dell XPS 720 H2C, when I get it in the next 2 months.Keep up the good work my friend!