Phenom 2.6 GHz. 140W. Late Q1 2008

[red]

http://www.extremetech.com/article2/0,1697,2218304,00.asp

Phenom 9900 2.6 GHz 140W TBD Late Q1 2008

Previously spotted over a month ago http://blog.livedoor.jp/amd646464/archives/51075170.html but it's from a better source now.

http://www.hothardware.com/articleimages/item1057/big_phenom_slide_5.jpg
PS, what does 110A vs 95A mean?

[Pete]

A is AMPS

110 & 95 explaine them self!

[red]

I know that. Is that all you can say?
http://www.legitreviews.com/article/597/10/
BTW, here's the 9900's 140W TDP in action.

[alucasa]

140w...

It looks like K10 needs to meet 45nm.

[Richard Dower]

Not intertested in AMD chips i must say, 45nm Intel seems to be the way to go...it's gonna take a long time for AMD to sort themselves out and get the steppings and revisions right, i bet we'll see a 90nm Conroe situation...with multiple chip revisions.

[JumpingJack]

140w...

It looks like K10 needs to meet 45nm.

Don't expect miracles, AMD thickened up the gate oxide for Barcleona/Agena and it didn't help much.... 45 nm, with SiO2 in the first incarnation, will not give back as much on power as you would hope for.

[BeardyMan]

i wonder what both companies take for reference as a test to test this?:)

I don't think they use 06 or such eh :p:

[alucasa]

Not expecting anything at all at the moment.

I've already had my share of nightmares while obtaining Opty 2346HE. My next cruncher will be Clover.

[Lightman]

http://www.extremetech.com/article2/0,1697,2218304,00.asp

Phenom 9900 2.6 GHz 140W TBD Late Q1 2008

Previously spotted over a month ago http://blog.livedoor.jp/amd646464/archives/51075170.html but it's from a better source now.

http://www.hothardware.com/articleimages/item1057/big_phenom_slide_5.jpg
PS, what does 110A vs 95A mean?

That means nothing without second variable which is V(olts)!

I'm gonna ask what is worse:
1 - 95A at 1.3V
2 - 110A at 1.1V
??

PS. both are under 125W limit....

[Omastar]

Wow, that's a higher TDP than Smithfield!

Awesome.

[terrace215]

Don't expect miracles, AMD thickened up the gate oxide for Barcleona/Agena and it didn't help much.... 45 nm, with SiO2 in the first incarnation, will not give back as much on power as you would hope for.

Without HK/MG, AMD's 45nm might use *more* power than 65nm at a given clock. Leakage will be horrendous.

[Jamesrt2004]

Major Leakage ???

the difference between idle n' load is HUGE

[Nedjo]

Wow, that's a higher TDP than Smithfield!

Awesome.
LOL they've pulled 140W out of they ass and you've took it for granted!

Just 'cos sAM2+ and AM3 requests 110A PWM it doesn't mean that next iteration of Phenom will suck up 140W!

[Omastar]

LOL they've pulled 140W out of they ass and you've took it for granted!

Just 'cos sAM2+ and AM3 requests 110A PWM it doesn't mean that next iteration of Phenom will suck up 140W!

Judging by the current TDPs and leakage woes experienced by Phenom, this 140W TDP actually sounds reasonable to me. Even if it's not, let's say 130W. That's still on par with a Smithfield's nuclear-grade power usage.

And granted, Intel's QX9770 also runs incredibly hot according to Anand, but that's not final production silicon. B2 Phenoms and B3 Phenoms aren't going to be any world apart, and I doubt either will scale very well with more volts.

[knightwolf654]

Don't expect miracles, AMD thickened up the gate oxide for Barcleona/Agena and it didn't help much.... 45 nm, with SiO2 in the first incarnation, will not give back as much on power as you would hope for.

is there a reason AMD wont go with that material intel is using?

[STaRGaZeR]

This... just... sucks...

I´m waiting for very affordable Phenoms, so Intel will be forced to do a price war with it´s new Yorkfield. Damn I want one of these 45nm parts NOW!

[Ub3r-L33ch]

AMD is dead to me ...



J/K



But damn, they need to hurry up and start scaring Intel again so we can have some good old price/performance wars :D

[Shintai]

is there a reason AMD wont go with that material intel is using?

SOI, arrogance (Native quad, not dumping SOI etc, Hector seems the key issue here) and their IBM partnership.

IBM is maybe not even in top5 anymore...they have a bad burden of going exotic without longterm thinking.

If AMD is still trying SOI for 45nm...it might not be Q1 2009 for 45nm. But lucky if in 2009 at all.

[JumpingJack]

is there a reason AMD wont go with that material intel is using?

High-K has been in research/lab phase for the past 10 years, if you backtrack through the IEEE journals, and conferences ... it was clear Intel had a head start in tackling the integration problems. There are a hand full of issues with high-K integration from Fermi Pinning to breakdown reliability.

So ultimately I expect AMD and the entire industry to move to high-K, both IBM/AMD has stated that they intend to do this relatively soon, AMD in their July technology meeting were not firm ... stating in a 2nd rev of 45 nm or 32 nm.

Earlier, before Intel disclosed thier high-K intentions, the fab 5 group (IBM, Tosh, AMD etc) had stated publically that high-K would intersect the 32 nm node... I think Intel's revelation caught them off guard.

Who knows... the only thing that is clear from AMD is that the first revision of the 45 nm process will use SiO2. That has made it into most of their published presentations.

If there is one reason why AMD will be behind on this tech is simply because they have not engineered a means to mass produce it yet, Intel simply is ahead.

Jack

[Piotrsama]

SOI, arrogance (Native quad, not dumping SOI etc, Hector seems the key issue here) and their IBM partnership.

IBM is maybe not even in top5 anymore...they have a bad burden of going exotic without longterm thinking.

If AMD is still trying SOI for 45nm...it might not be Q1 2009 for 45nm. But lucky if in 2009 at all.

I was thinking the same: Is SOI paying off?
On the contrary, looks like a burden.

[Zytek_Fan]

SOI, arrogance (Native quad, not dumping SOI etc, Hector seems the key issue here) and their IBM partnership.

IBM is maybe not even in top5 anymore...they have a bad burden of going exotic without longterm thinking.

If AMD is still trying SOI for 45nm...it might not be Q1 2009 for 45nm. But lucky if in 2009 at all.

And Hector will still be able to do more damage, since he doesn't leave until April :(

F**K YOU HECTOR RUIZ!!!!

[JumpingJack]

I was thinking the same: Is SOI paying off?
On the contrary, looks like a burden.

SOI provides a handful of benefits, what you won't hear from AMD/IBM are the disadvantages... self-heating, kink-effects, exacerbated HCI etc.

The benefit of SOI as you scale down decreases to the point that Intel determined at 65 nm was less than 5% over bulk in the parametrics it was suppose to help.

http://www.intel.com/technology/silicon/ieee/soi2000.pdf

You can see this particularly strongly in PMOS (figure 5) and the FO4 delay (figure 11/12).

Essentially, as one goes smaller on PD-SOI the bang that it brings diminishes...

Jack

[Wiggy McShades]

if it weren't for the power consumption they would of been decent chips if priced effectively. unfortunately amd is still doing an almost job yet again. rv670 is almost as good as its competition and phenom is almost there in terms of a product with a decent value.

[Mortal]

I wonder if AMD recently hired a guy called Murphy? :shrug:

[Dimitriman]

No thanks.. waiting for Fusion now =[

[Zytek_Fan]

Forgot to say:

This is bad.

[saaya]

LATE q1? meaning almost q2? :slapass:

[BrowncoatGR]

They need SGOI asap actually. I hope they are still on track to get the first chips out soon.

[KeZzZu]

Actually this is fun, reminds me somehow my cabge 0536 oppy (144)
most of them maxed out at 2.5-2.8 , well i just 24/7 it with 1.8vcore and 2.8ghz :D max air were somewhere near 3ghz

So bring it on amd again, some fun time outside house, you get decent beach on your yard thought?? with this power usage :P

[RaZz!]

when i first saw the graphs i was just like oh, phenom at 2.6ghz is better than intel's qx6850, but then i realized that the graphs show power consumption. damnit lol

[Shintai]

So bring it on amd again, some fun time outside house, you get decent beach on your yard thought?? with this power usage :P

LOL, Stop global warming!! :p:

On a sidenote AMD is getting the short end of the stick on the stock market now. Down to about 11.5 and dropping. Maybe if Hector gets a heart attack before he stops....

[OBR]

After AMD Warsaw talking ... is situation very bad for whole company ... AMD dont have all money they will need to make a B3 revisons of their processors ... They are in END of their Days ... AMD needs big help from new investors, last money injection from Abu Dabbi keep AMD live for some months, but after that AMD will die!

[deathman20]

The high power consumption is a killer. Sure Intel's was high at first as well but still it was lower than that. Sucks though first it was suppose to compete with Intel's quads, Failed... then was suppose to be low powered, Failed... now low priced for the performance, we'll really have to wait and see how that plays out.

[freeloader]

SOI, arrogance (Native quad, not dumping SOI etc, Hector seems the key issue here) and their IBM partnership.

IBM is maybe not even in top5 anymore...they have a bad burden of going exotic without longterm thinking.

If AMD is still trying SOI for 45nm...it might not be Q1 2009 for 45nm. But lucky if in 2009 at all.

I usually don't agree with what you say, but in this case you're 100% right. AMD really needs to ditch SOI and fast.

[Brother Esau]

LATE q1? meaning almost q2? :slapass:

As much as I hate to say it that may be a tad optimistic considering,

[GoThr3k]

LOL, Stop global warming!! :p:



stop folding :P
J/K

SOI has its benefits though,if only Z-ram would work at 2ghz :)
too bad SOI counters the leakage that isnt important anymore

[Shintai]

stop folding :P
J/K

SOI has its benefits though,if only Z-ram would work at 2ghz :)
too bad SOI counters the leakage that isnt important anymore

Heh, ZRAM...thats still funny. ZRAM was never and will never be intended to performance CPUs. It was targetted at embedded CPUs. Its been a hype ever since.

SOI had benefits at 130 and 90nm. It only got disadvantages at 65nm and down.

[G80]

My 130nm fx 55 draws less then that.

[Slay0r]

My 130nm fx 55 draws less then that.

Well you'd expect a single core to draw less power than a quad :p:

[G80]

Not with twice the die size and no power management whatsoever.

[terrace215]

http://img518.imageshack.us/img518/7876/amdphenomturkeylogoeg9.jpg

[vitaminc]

I usually don't agree with what you say, but in this case you're 100% right. AMD really needs to ditch SOI and fast.

I still remember back in the earlier part of 2006 people were still thinking SOI is AMD's saving grace for low power consumption.

[The_Beast]

I'm not a fan of AMD but they need to kick it in gear