NRP said:
Alright, I installed the P4c 3.4 GHz cpu, and here are some comparative benchmarks:
P4c 2.6 GHz:
Doom 3 (1024x768, Med): 56.4
Doom 3 (1600x1200, High): 47.4
Halo PC (1600x1200, High): 54.12
3DMark03 (default): 11,852
P4c 3.4 GHz:
Doom 3 (1024x768, Med): 66.6
Doom 3 (1600x1200, High): 51.9
Halo PC (1600x1200, High): 60.05
3DMark03 (default): 12,491
Nice increases, actual gameplay is noticeably improved. However, I was kind of expecting more . . .
so . . .
I'm building a new system based on an AMD64 FX-53 (socket 939) cpu and an ASUS A8V Deluxe motherboard. At this point in my life, I've got more dollars than sense, but I just want to see how much of a difference the fastest cpu/motherboard combo currently available will make with my videocard. Everyone says the X800 XT PE needs a fast cpu to really perform. We'll soon see.
I've got the new system up and running, and here are the results as well as some conclusions:
AMD64 FX-53 (Socket 939 @ 2.4 GHz) w/1 GB PC3200 RAM:
Doom 3 (1024x768, Med): 91.5*
Doom 3 (1600x1200, High): 56.8
Halo PC (1600x1200, High): 67.15
3DMark03 (default): 12,827
*Since Doom3 is cpu-limited at low resolutions, I think this benchmark score is most indicative of how much faster the FX-53 processor is compared to the P4c 3.4 GHz. All of the other benchmarks are basically evaluating the graphics card.
As soon as I got this thing running I immediately ran benchmarks, and, admittedly, I was initially disappointed by the results. These benchmarks seem to show that it isn't worth dropping ~$1,500 (US) on a new FX-53 based system. I guess I was expecting larger benchmark score differentials.
Then I started to play games. Three words best describe the new system: Hol-E-Crap! The actual gameplay feels like a "night-and-day" difference. Basically, the framerate dips are much smaller now, even at 1600x1200 resolution with the highest settings. I guess when you throw in Havoc physics and destructable environments (I love shooters), the FX-53 really shines.
What did I learn from all of this?
1. Well, benchmarks don't tell the whole story (everyone always says this, and I just spent a lot of money to learn it's true).
2. Above a certain minimum specification (a 2.6 GHz P4c cpu in my case), a good video card is more important than a good cpu.
3. The best benchmark is to play your favorite games. If your current system experiences large variations in in-game framerates, then upgrade your videocard first, cpu second. Unlike cpus, the latest videocards are so much faster than the last generation it's ridiculous.
I know I'm not really telling you people anything you don't already know, but if you're like me, you don't necessarily believe everything you read. I actually spent the money and did the research. I hope this experiment will be helpful to someone.
Oh yeah,
4. Far Cry and Painkiller are a hellava lot more fun to play than Doom3.