A64 3000 overclocked vs A64 3800

[lenjack]

I was all set to get an A64 3800+ and overclock it...I know it will only o/c to a limited extent, but I see Anand's test showing the 90 nm A64 3000+ overclocking to 2.6G. This would save a lot of $. How does performance of an overclocked 90 nm 3000 compare to an overclocked 3800?

 

[Markfw]

There was a lot of people that say it can be done most of the time, but I doubt it. I am having trouble getting my 3200+ up to 2350 stable ! Its still a guess, but the 3800+ doesn;t have be be OC'ed much....

 

[jpeyton]

To be fair, some people who have bought 90nm have had the same amount of success as Anand, some haven't.

And since overclocking isn't a guarantee, we don't know how high a 3000+ and a 3800+ will go.

At identical clock and bus speeds, they perform almost indentically (the 90nm cores perform slight better).

But if the clock speeds are the same and bus speeds aren't, the faster bus speed will win. And vice versa.

 

[lenjack]

Cool guys...anyone else?

 

[Rhagz]

The cpu that allows a higher multiplier generally can oc a little better imo. You may get a 3000 that can run 300FSB all day long, but odd are you wont. If you got one that handled 250ish, and only a multiplier of 9, then your limit is gonna be 2250 no matter what you do. On the 3800 however, if you only got a FSB of 250, but a multi of 12, then you could concievably hit 3k. Most people say that higher FBS and lower multi is better, but not much DDR can handle 600mhz heh, then you will be limited there as well. If you get ram that runs good at 233 mhz, then having those extra multipliers helps alot. 233x12=2796, but 233x9 is only 2097.. so you can clearly see where a higher multi can be beneficial.

 

[Mrvile]

OCing a 1.8 3000+ to 2.6 is as high as they go. Most people can't get higher than 2.4, and if you manage 2.6 well be careful you don't hurt yourself. Or your CPU.