• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

New AMD roadmap sucks

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
xtrmeocr said:
uhg do you read or what?

Why yes I do, thank you.

xtrmeocr said:
the entire line would be all the chips based on that chip revision. we are at 2.4ghz now and i think there will be a 2.6ghz 130nm version, i think its safe to say 90nm will bring us to atleast 3ghz.

How can you be so sure 90nm will bring AMD to 3GHz? Like many people have noted, the power consumption just takes off at some point and ceases to be linear with clock speed. The same can be said of Intel and IBM. Intel cannot reach 4GHz with Prescott and IBM cannot reach 3GHz with PowerPC G5.

xtrmeocr said:
good for them how does that have any effect on amd?

Because it is a good foreshadowing of the problems that AMD is having with the 90nm process. Intel can hardly get any 3.4GHz Prescotts out the door, even months after announcing them. AMD is partners with IBM, and IBM is having problems with 90nm. AMD's investment in SOI is not a magical cure.

xtrmeocr said:
what are you trying to say here?

I'm trying to say that Intel is not able to keep their power consumption the same while scaling up. If they were able to, they'd be up at 4GHz already. AMD could have the same problems.
 
you cannot compare different marchitectures and processes to each other.

yea ibm hasn't been able to produce a 3ghz ppc, apprently they dont have anything over a 2ghz ppc, however AMD is producing 2.4ghz athlon64s and Opterons comfortably, with 2.6's to come.

intel is even farther off the mark because they are using strained silicon and low-k dielectrics.

of course they cannot keep the power consumption the same at each speed grade, the speed is part of the power consumption equation.
 
Silent Buddha said:


Your post is wrong. That's not how the Socket 939 PR system works. Socket 939 A64 3500+ is clocked @ 2.2GHz and the Socket 939 A64 3800+ is clocked @ 2.4GHz. Now tell me how a Socket 939 A64 4200+ extrapolates to 3GHz? And why would that make sense? A64 FX-55 by itself requires 104 watts/80 amps. The FX-55 is clocked @ 2.6GHz and needs 104 watts so why would a 3GHz A64 only require 104 watts?
Its not wrong its speculation.
I am speculating that it will only be 512KB for L2 cache and as my post says at present a 2Ghz is 3000+ and a 2.2GHz is 3200+ add an extra 200+ for the one-off effect of 939 dual channel and extrapolate up at 200Mhz = 200+ points and you get 4200+ = 3Ghz.
The 3500+ 939 being 2.2Ghz and 3800+ being 2.4Ghz is also similar speculation (though I wouldnt say it is necessarily wrong). The problem with that is for the 2.2GHz compared to a 754 the difference is 300+ points (3500 v 3200) but for the 3800+ the difference mysteriously increases to 400+(3800 v 3400).
Truth is nobody yet (probably including AMD!) knows how they are going to do their rating system on 939.
 
OC Detective said:

Its not wrong its speculation.
I am speculating that it will only be 512KB for L2 cache and as my post says at present a 2Ghz is 3000+ and a 2.2GHz is 3200+ add an extra 200+ for the one-off effect of 939 dual channel and extrapolate up at 200Mhz = 200+ points and you get 4200+ = 3Ghz.
The 3500+ 939 being 2.2Ghz and 3800+ being 2.4Ghz is also similar speculation (though I wouldnt say it is necessarily wrong). The problem with that is for the 2.2GHz compared to a 754 the difference is 300+ points (3500 v 3200) but for the 3800+ the difference mysteriously increases to 400+(3800 v 3400).
Truth is nobody yet (probably including AMD!) knows how they are going to do their rating system on 939.

Yeah, I take back what I said. I guess we cannot be for sure. I got my info off one of the Anandtech roadmaps :-/
 
Back