So how are those Barton's doing? Overclocks? Performance?

[MichaelD]

Was just browsing around and saw an article on the Bartons. They've been out awhile, but haven't seemed (IMO) to have the impact of say, those overclockable XP2100s that everyone got. (me too!)

AFAIK, the Barton has 512kb of L2 cache (I think it's L2) and a default speed of 166MHz. That's it, right?

How are they overclocking? Anything like the 2.1GHz I get out of my XP2100?

 

[bgeh]

they usually hit 2.2-2.3GHz on 1.85v
and yes, the Barton has 512kb of L2 cache
its fsb is 166

 

[HappyCracker]

My 2500 hit 2250 at 1.8v. Stock cooler at that (Delta 38 on top though). Seriously, I think that fan would make anything a decent cooler. I digress, You probably won't be hitting the 2400 MHz I've seen people with on their 1700+, unless you've got badass cooling like VapoChill or Prometia. I read some guy on a VC got 2600 MHz out of his. If only they cooled more stuff like the vid card and chipset, those things would be sick. Rambling.....END

 

[MichaelD]

Cool. Thanks guys.

I guess the Bartons aren't all that popular w/the enthusiast crowd, huh?

We're all waiting for Opterons (sp?) huh?

 

[mrman3k]

Bartons don't overclock as high as the Tbreds. As for Opterons, don't expect anything as far as overclocking, they are pretty much the same as current Athlon XP's with a bit more cache, etc.

 

[MadRat]

FSB probably isn't as necessary for the Barton cores. Perhaps people should try to scale back the FSB and play with the larger multipliers?

 

[HappyCracker]

While you could probably get a higher multiplier, it would hurt overall system performance. Given DDR400 and running the FSB at 200 (400DDR), you would probably see better performance a cross the board due to the fact that the system can talk to its components faster. This is also why AMD kicked the FSB up to 400, it's a known way to yup, you guessed it, performance. Same goes for Intel and their 800 (200 QDR) bus speed they just instantiated. Reading something somewhere (red hill page linked off of sandpile.org i believe) they show (or say rather) that the systems with higher bus speeds win out in the long run because the components attached to the processor can keep it fed with that good old data lovin'