StarDestroyer.Net BBS

Anandtech.com linky

Amazing, we're talking anywhere from 20% improvement to 40%.

Lets try hotlinking them.

And the chip test I read a couple weeks ago had AMD spanking Intel in every catagory, especialy games. And they were testing the highest of the high end.

Alyeska wrote:And the chip test I read a couple weeks ago had AMD spanking Intel in every catagory, especialy games. And they were testing the highest of the high end.

Chip test where? With what CPUs?

Let me guess, a currently out Pentium D?

Let me clarify: Conroe is not out yet, Conroe is pre-production sillicon to be released at the middle of this year, with a completly new architecture(yes, I know its not 100% new).

Alyeska wrote:And the chip test I read a couple weeks ago had AMD spanking Intel in every catagory, especialy games. And they were testing the highest of the high end.

They were probably testing high-end Pentium D/EE processors; this appears to be the first public Conroe benchmark but under blackbox conditions. Benchmark cooking is a long tradition but 20% gain over K8 is very reasonable (I'd be shocked if Intel didn't deliver on those gains).

Ace Pace wrote:Let me clarify: Conroe is not out yet, Conroe is pre-production sillicon to be released at the middle of this year, with a completly new architecture(yes, I know its not 100% new).

Intel Core Microarchitecture (P7?) is supposedly an entirely new design but strongly influenced from the Pentium M line.

phongn wrote: They were probably testing high-end Pentium D/EE processors; this appears to be the first public Conroe benchmark but under blackbox conditions. Benchmark cooking is a long tradition but 20% gain over K8 is very reasonable (I'd be shocked if Intel didn't deliver on those gains).

Wouldn't the fact that these X2 benchs match previously recorded numbers taken from the X2 pretty much deny cooking?

Intel Core Microarchitecture (P7?) is supposedly an entirely new design but strongly influenced from the Pentium M line.

Yes, influenced actully by both the Pentium M and the Netburst, from what I understand they are trying to take the best of both worlds.

Ace Pace wrote:Wouldn't the fact that these X2 benchs match previously recorded numbers taken from the X2 pretty much deny cooking?

Intel could have done some tuning voodoo with Conroe one way or another.

Yes, influenced actully by both the Pentium M and the Netburst, from what I understand they are trying to take the best of both worlds.

Yes, but there are other features of it not seen in either P6 or P4.

phongn wrote:

Ace Pace wrote:Wouldn't the fact that these X2 benchs match previously recorded numbers taken from the X2 pretty much deny cooking?

Intel could have done some tuning voodoo with Conroe one way or another.

How would they go around doing that?

I wonder how AMD's next big thing stacks up against Conroe.

AnandTech has some more details about Conroe in their Spring IDF 2006 article

Ace Pace wrote:How would they go around doing that?

No real clue, honestly

Arrow wrote:I wonder how AMD's next big thing stacks up against Conroe.

K8 is moving to DDR2, which has greater bandwidth but also greater latency. No major core revisions are planned for the near future, AFAIK.