FX80 and FX82 4 and 4.4GHZ quads from AMD?

[Tony]

Source: http://reviewage.co.uk/content/view/33/1/

Could this be true? I know AMD keep very quiet so no one has heard a peep from them of late.

There's said to be two chips in the pipeline, the FX-80 at 4Ghz and the FX-82 at 4.4GHz.
The FX-80 at stock out performs a 5.0GHz Kentsfield.

[Jochenp]

http://www.fudzilla.com/index.php?op...2984&Itemid=35
A Fudzilla link, that says it's to be released late Oktober of 2007, but at speeds of 2Ghz and 2.4Ghz. But yeah, that's the Fud ofcourse...

[Marvin_The_Martian]

That was posted 4 days ago, and no repleis and no one picked it up besides you. Well maybe they did but you're the first I heard this from.

I'm not believing anything untill I see it. Though maybe the reason Amd hasn't made some waves about this is, that they know it will go up against Nehalem not Kentsfield

[003]

If it's true, very impressive. But I'm not holding my breath. A 4.4GHz stock speed on any Phenom is a little bit hard to swallow, ya know?

[Solarfall]

i'll belive it when i see it...very impressive if this news turns out to be true

[Tony]

http://www.fudzilla.com/index.php?op...2984&Itemid=35
A Fudzilla link, that says it's to be released late Oktober of 2007, but at speeds of 2Ghz and 2.4Ghz. But yeah, that's the Fud ofcourse...

thats long gone, so forget Fudo's info this time

[DMH]

Yeah right,from the slow 65nm they go to a super quick 45nm that beats 5Ghz Yorkfield

[BulldogPO]

I´m very, very doubtful but time will tell.

[lm358]

Hmm FX80 4GHz > 5GHz Kentsfield... well thats very impressive! And I can fully believe it, I mean surely AMD has known for over a year before they even launched Phenom they needed to make some serious improvements to catch back up (and overtake) Intel in performance per clock. I hope its a RV770 like resurgence, would love to switch back to AMD

[Kunaak]

AMD has leap frogged Intel Before... remember the AMD64? took intel 4 years to come up with something decent enough to take the spotlight away from it.

they did it before. they can do it again.
its happend in the videocard market a few times too.
most notably with the 9700 pro.

[Solus Corvus]

It would be nice if true. But it seems totally unrealistic based on our current information.

Though they did pull off the unexpected with RV770. Could their processor department pull off the same sort of coup? Possibly, but wouldn't we have heard a little something, even vague rumors, before now?

[duploxxx]

oh boy oh boy, this will be a nice flame thread again.... the deneb fx exists in the roadmap but without speedgrade and final release date. so that will be q1-q2 2009.

[AliG]

I hope this is true, especially since phenom's performance scaling is better than linear in a lot of cases as frequency goes up due to the bottlenecks becoming eliminated/less, but I just can't picture amd going from 2.6ghz to 4ghz (stock) with a simply die shrink unless if shanghai is much more than just a faster ipc and die shrink

[Nedjo]

This reminds me on some other similar "announcements" from some other sites with similar reputation and presumably same low traffic! This will definitely bring some high speed rotation of their traffic counter regardless of the fact it's utter FUD!

[Macadamia]

FUD.

Anyway, I'm stuck on my camera purchase now. NOTHING beats a DSLR that can record HD. Not even Core i7 mighty holyness.

[SimBy]

Yeah right.

[Tony]

consider the following:

AMD since buying ATI have gone even quieter than they used to be, NOTHING leaks out of AMD, so what you think you know you do not, they hold secrets real tight, so there could be some truth to this.

We all know Phenom had to speed up. AMD have publicly said the same many times. So did they go back to the design table and rethink Phenom and inject some speed?

I doubt even this new FX if true would match Nahelem even though they are so related its untrue, QPI= HT and don't let anyone say its not, I bet they just decided to change the name to distance themselves from AMD.

I have seen similar jumps in yield on memory IC's to what they are saying here, so its not impossible.

[Hornet331]

god here we go again.... incredible bold statesments only to be disappointed.

What we have so seen so far deneb only catches up with Kentsfield, so the statement 4,4ghz deneb = 5 ghz kentsfield is just plain bs. Of course it depends on what pllication you use, i think that statement would be true for server loads just like spec_rate benchmarks, but how that translates to realwold, noone knows.

Also 4,4ghz stock frequenzy is another bold statement, going from 2,6 ghz to 4,4ghz... hard to belive....

[Macadamia]

Tony, if this is true, it adheres to what I've seen on Overclock.net. Aka The 4Ghz snapshot.

They must have changed logic. Perhaps they were using more static logic which uses less power at default but scales less impressively, then switched to dynamic logic because 45nm's low-k and shrink nature took power consumption down already.


Which is ironic because Nehalem is going in reverse. Dynamic (Penryn) to static (i7). But it works for Intel so don't flamestart.

[AliG]

well k10's ipc increases as the clockspeed increases due to various bottlenecks at the lower speeds, but eitherway it's not AMD saying this, it's some no-name site no one had heard about before this that could be faking the cpuz or just taking it from someone that had already demoed a 4ghz phenom that was oc'd to get there. So don't be disappointed if this doesn't happen as AMD never made this claim

[mAJORD]

I don't think so.

[clonez]

it sure is possible, those jumps happened a lot in early cpu times, and nowadays with gpu's

but i dont see how they should have done this, there have to be some mayor changes (high k anyone?)

i guess too goood to be true

edit: btw, they would bring a lot more processors than two fx, probably would ditch the whole 9xxx and 8xxx series

[Tony]

Tony, if this is true, it adheres to what I've seen on Overclock.net. Aka The 4Ghz snapshot.

They must have changed logic. Perhaps they were using more static logic which uses less power at default but scales less impressively, then switched to dynamic logic because 45nm's low-k and shrink nature took power consumption down already.


Which is ironic because Nehalem is going in reverse. Dynamic (Penryn) to static (i7). But it works for Intel so don't flamestart.

Now this is an educated guess, great post

I have a feeling if this is true they have definitely changed Phenom in some way. All along they have said all Phenom needed was more clock speed, this could be that clock speed

[clonez]

it most certainly does, scales much better than intels proc (though that can change with nehalem)

[mcbalaban]

I'm an AMD fan(boy) and I hate to state the obvious....

"The Deneb core apparently has a multiplyer of up to 25x!"

Guess what is the max multi on my ol' FX-57...?

...Yep. 25.

Not really a confidence-inspiring website, it seems...

Anyway, I'm sure that as soon as AMD releases such a CPU, I'll wake up and have to go to work on a cold, rainy Tuesday morning...