I've been a big fan of AMD systems for many years now. Great value for money. But it turns out the Intel Conroe fanboys aren't fanboys... they're just right! :-)
System (1):
AMD x2 3800 overclocked to 2.4GHz, 2Gb DDR, Asrock dual SATA, 8800GTS
System (2):
Intel e6600 @ 2.4 GHz, 2Gb Crucial Ballistix (5300 but famous for working well beyond that at a bargain price), Gigabyte DS3, 8800GTS.
System (3): as system (2) but CPU overclocked to 3.15GHz (I haven't tried hard to push it any further yet, as I have a recent e6600 which needs higher voltage and is not quite as overclockable as earlier CPUs).
LOMAC test (a):
An F2 view from behind an Su-25 flying low over Sevastopol.
LOMAC test (b):
A short 25T mission involving a rocket attack (lots of smoke & explosions) on a base defended by anti-aircraft missiles etc.
Fraps is used to measure the average, highest, and lowest FPS in 60 second replays of tracks from missions (a) & (b). LOMAC was run at 1280x1024, x4fsaa, x8aniso, *high* visibility range, medium water, and everything else at maximum apart from shadows, due to 8800 shader problems.
______________(1)______(2)_______(3)
(a) AVERAGE___10.2______18_______23
___MIN FPS_____8_______15_______18
___MAX FPS____13_______23_______30
(b) AVERAGE___21.8______30.5_____37.4
___MIN_________3_______14_______12
___MAX________35_______46_______54
I could throw more numbers at you, but I think those LOMAC figures say it all. They're not scientific tests, averaged over many repetitions, and the minimum FPS can be very variable. But those results tie in very well with what I've seen in more "standardised" benchmarks like Aquamark3 and the various 3dmarks. And LOMAC really does look and feel smoother, especially during intense action, and over cities. I was expecting to be disappointed, to feel I'd wasted even *more* money chasing better LOMAC performance... but this upgrade was definitely worth the money, even though I've only had the x2 3800 a year and I usually make a processor last two years or more.
Hopefully these numbers will be useful to someone else trying to justify an upgrade.
Andrew McP
PS I did a quick Vegas (Movie Studio) render test. On my x2 3800 it took about 1 minute 30s to render a 9 second test clip with various effects added. On the e6600 setup @ 3.15GHz it takes 36 seconds. That 4Mb of dynamically shared on-chip cache really seems to make a huge difference compared to the 2x512k on my AMD. Of course owners of better (more cache/faster) AMD CPUs would see less of an improvement.