Friday, July 10th 2009

AMD Staring at 140W Barrier with Phenom II X4 965?

Two of AMD's biggest setbacks with the 65 nm Phenom X4 series were 1. the TLB erratum fiasco with the B2 revision of the chip, and 2. the virtual TDP wall it hit with the 2.60 GHz Phenom X4 9950, at 140W. At that wattage, several motherboards were rendered incompatible with the processor because they lacked the power circuitry that could handle it. The company eventually worked out a lower-wattage 125W variant of the said chip, and went on to never release a higher-clocked processor based on the core.

MSI published the complete CPU support list of its a new BIOS for the 790GX-G65 motherboard a little early, revealing quite some about unreleased AMD processors. At the bottom of the list its the Phenom II X4 965. This 3.40 GHz quad-core chip will succeed the Phenom II X4 955 as AMD next flagship desktop offering. Its TDP is an alarming 140W. Alarming, because this is a chip with a mere 2 unit bus multiplier increment over the Phenom II X4 940, the launch-vehicle for AMD's 45 nm client processor lineup. There are, however, two things to cheer about. RB-C2 is not going to be the only revision of this core, future revisions could bring TDP down, or at least make sure clock-speeds of future models keep escalating, while respecting the 140W mark. A future variant of Phenom II 965 could come with a reduced TDP rating. The list interestingly also goes on to reveal that AMD will have a 95W version of the 3.00 GHz Phenom II X4 945.
Source: HardwareLuxx.de
Add your own comment

184 Comments on AMD Staring at 140W Barrier with Phenom II X4 965?

#76
FordGT90Concept
"I go fast!1!11!1!"
DippyskoodlezAn AXP 2600+ was rated the same TDP was the 3200+.
TDPs are generally measured and labeled for an entire line, not individually. To be safe, they usually use the highest spec'd processor to set it. In your example, the TDP on the 2600+ was probably what was tested on the 3200+. They then design the HSF and motherboards to power/cool the 3200+ knowing that any lower model of processor will be perfectly safe.

The same can be said of modern CPUs. X58 motherboards and HSF are designed to power and cool a Core i7 965. That way, they can use the same motherboards and HSFs to run 950s, 940s, and 920s. The specification may have to be changed for Core i7 975 and then again, might not. It depends on changes between the two and whether or not they over rated the TDP on the Core i7 965.
Posted on Reply
#77
Dippyskoodlez
FordGT90ConceptTDPs are generally measured and labeled for an entire line, not individually. To be safe, they usually use the highest spec'd processor to set it. In your example, the TDP on the 2600+ was probably what was tested on the 3200+. They then design the HSF and motherboards to power/cool the 3200+ knowing that any lower model of processor will be perfectly safe.
Rewording my post? :wtf:
Posted on Reply
#79
Kitkat
Darrenmtosev,

The CPU has not been released yet, all this information is based on sniffing information, wait for to see when or if the CPU gets released at 140w before we call them a fail, secondly what has what has MHz got to do with anything?

Has anyone ever noticed, its always the guys with ancient processors such as a the Pentium 4 *cough cough* or Celeron or something stupidly slow and old that always bash AMDs flagship processor. They would cream in their pants to swap their Pentium 4 for a Phenom II.
rofl yep Mr Wtosev Presscott (lmao woops)
TheMailMan78Honestly I was expecting more out of a 140w chip. Kinda a let down from AMD.
Thats why say 975 will have the lower twp they already lowered it but i think itll be in that one. 965 looks like bump from this info. A while back there was a story about it but i think there still testing it?
Posted on Reply
#80
cdawall
where the hell are my stars
just a curiosity of mine why are these chips being compared to prescotts? prescotts were known for being absurdly hot while not improving at all over previous generation chips. phenom II is for one stock clocked@3.4ghz which is higher than damn near every 65nm K10 chip will clock and runs very cool even at 3.8ghz these chips put out less heat than a prescott single core.


as for the 140W thing that is a max TDP most of these chips will never hit that. hell my 65nm 9150e downstairs is rated at 65w TDP it doesn't even pull 40w right now.
Posted on Reply
#81
Kitkat
cdawalljust a curiosity of mine why are these chips being compared to prescotts? prescotts were known for being absurdly hot while not improving at all over previous generation chips. phenom II is for one stock clocked@3.4ghz which is higher than damn near every 65nm K10 chip will clock and runs very cool even at 3.8ghz these chips put out less heat than a prescott single core.


as for the 140W thing that is a max TDP most of these chips will never hit that. hell my 65nm 9150e downstairs is rated at 65w TDP it doesn't even pull 40w right now.
meh quote of a quote of a quote ,the guy arguing "amds fail" ironicly has presscott. And the original presscot comment was wayyy off ;) hope that clears it up for u lol
Posted on Reply
#82
cdawall
where the hell are my stars
Kitkatmeh quote of a quote of a quote ,the guy arguing "amds fail" ironicly has presscott. And the original presscot comment was wayyy off ;) hope that clears it up for u lol
no idea why this is huge 140w is meh most people overclock to 3.8ghz @1.4v on the 955's which pulls like 200w+
Posted on Reply
#83
TheMailMan78
Big Member
cdawallno idea why this is huge 140w is meh most people overclock to 3.8ghz @1.4v on the 955's which pulls like 200w+
I'm on 1.44v right now ;)
Posted on Reply
#84
Flyordie
btarunrUh what? All K10 CPUs have two memory controllers, but those are two independent 64-bit memory controllers (hence the ganged/unganged DCT modes). It's not that one is DDR2 and the other DDR3. The IMCs on Phenom II AM3 chips support both DDR3 and DDR2, it's not that there are two sets of memory controllers based on the standard.
After looking at the engineering notes, my idea was correct but it was backwards.
Its a DDR2 controller with DDR3 "Extensions"... so maybe it won't work... they used alot of shared silicon on the PII's IMC.
Nothing that can't be changed through some creative engineering from the guys over in India. (the engineering team that developed Phenom II)
Which also leads me to say this- I just bought 2x Istanbuls for $18/each (shipping) direct from AMD as ES's with unlocked multi's to test out their maximum thermal load limits. Will be fun I guess destroying them... *cries*
Posted on Reply
#85
cdawall
where the hell are my stars
TheMailMan78I'm on 1.44v right now ;)
that puts you at ~170w which is about what my 550BE pulls@1.55v and 4ghz in perspective at the same setting a athlon II X2 250 pulls ~145w
Posted on Reply
#86
TheMailMan78
Big Member
cdawallthat puts you at ~170w which is about what my 550BE pulls@1.55v and 4ghz in perspective at the same setting a athlon II X2 250 pulls ~145w
1.5v seems a little high for 24/7 use. I would be careful with that.
Posted on Reply
#87
Flyordie
TheMailMan781.5v seems a little high for 24/7 use. I would be careful with that.
I think 3.4Ghz is the sweetspot for Deneb. 3.4 @ 1.35V is pretty good performance/watt.
Posted on Reply
#88
Kitkat
cdawallno idea why this is huge 140w is meh most people overclock to 3.8ghz @1.4v on the 955's which pulls like 200w+
me neither it was s far off base. my 955 is def over 140 rofl
Posted on Reply
#89
eidairaman1
The Exiled Airman
TheMailMan781.5v seems a little high for 24/7 use. I would be careful with that.
i dont see what the problem is, most had Athlon XPs at 1.9/2.0VCore when overclocked at 2.7-3.0GHz
Posted on Reply
#90
Kitkat
3xploit1. my 920 runs at 3.9ghz on air and loads at mid 60s - which is perfectly fine for any chip (amd or intel)
2. true, but i get much more performance
3. most people buy high end air coolers or water on these forums anyways regardless if they use intel or amd
4. i run my whole setup in a $50 antec 300
5. LOL ok there
6. no i7 x58 boards even use integrated graphics so wtf are you saying
7. core 2 quads still hold their own against phenom ii's and keep up with i7s in gaming

8. you are an amd fanboy lol
phenom II dosnt need to keep up with any f i7's "gaming" the only game i see is paying more for nothing. This isnt a video card discussion.
Posted on Reply
#91
Steevo
My 940 is running 1.55 vcore .05 higher than it is rated for, and AMD shows 125W at 1.5V


Considering at stock I can run just under a volt .975 including my vdroop it shold be labeled a 80W TDP chip, and I only need a couple extra tenths to make stress testing stable at this speed, if I back it off 100Mhz I can run 1.45vcore and it will be stable.


Personally I can run GTA4 at max res and high settings with no issues, and it didn't cost through the nose like a comparable Intel platform woudl have cost, and the 940 was a drop in replacement for my 9850BE, and I can run newer chips in this just fine, and the supposed loss from not having DDR3 is almost nonexistant. Plus the option to Xfire, with four cards.......
Posted on Reply
#92
mtosev
Kitkatrofl yep Mr Wtosev Presscott (lmao woops)



Thats why say 975 will have the lower twp they already lowered it but i think itll be in that one. 965 looks like bump from this info. A while back there was a story about it but i think there still testing it?
strange that that is my older pc. for which dont have any DDR ram anymore. hence it wasnt used in the past 2 years. to lazy to update the configuration tab. anyway my system now is a Core 2 Duo E6600, Asus P5w DH Deluxe, 2GB Gskill ram,...
Posted on Reply
#93
Steevo
mtosevstrange that that is my older pc. for which dont have any DDR ram anymore. hence it wasnt used in the past 2 years. to lazy to update the configuration tab. anyway my system now is a Core 2 Duo E6600, Asus P5w DH Deluxe, 2GB Gskill ram,...
$$$$ processor, and slower on average than a chip that costs 25% less
Posted on Reply
#94
btarunr
Editor & Senior Moderator
cdawallas for the 140W thing that is a max TDP most of these chips will never hit that. hell my 65nm 9150e downstairs is rated at 65w TDP it doesn't even pull 40w right now.
TDP is a company rating, not a measurement.
FlyordieAfter looking at the engineering notes, my idea was correct but it was backwards.
Its a DDR2 controller with DDR3 "Extensions"... so maybe it won't work... they used alot of shared silicon on the PII's IMC.
Nothing that can't be changed through some creative engineering from the guys over in India. (the engineering team that developed Phenom II)
Which also leads me to say this- I just bought 2x Istanbuls for $18/each (shipping) direct from AMD as ES's with unlocked multi's to test out their maximum thermal load limits. Will be fun I guess destroying them... *cries*
No, your idea was wrong all along. You said you could disable DDR2 controllers, and save power (insert absurd amount here). While I said there are two memory controllers which support both DDR2 and DDR3. You're getting technical on how I am right. The memory controller is designed that way, making it DDR3-exclusive is not going to cut its energy draw. Learn how things work.

Besides, AMD's market-share will plummet if they come up with AM3-exclusive chips. Nobody with decent AM2+ boards will continue using AMD because an upgrade-path end requiring you to buy a new board and memory. AMD retained a lot of market share banking on the backwards-compatibility of these processors.

Yes, Istanbuls are worthless at least for the client platform. No wonder they're giving away their ESes for scrap prices.
Posted on Reply
#95
mtosev
Steevo$$$$ processor, and slower on average than a chip that costs 25% less
yep but mine is now almost 3 years old. :)
Posted on Reply
#96
Wile E
Power User
cdawallno idea why this is huge 140w is meh most people overclock to 3.8ghz @1.4v on the 955's which pulls like 200w+
Because most people don't OC. 140w is most definitely an issue when it comes to OEMs.
eidairaman1i dont see what the problem is, most had Athlon XPs at 1.9/2.0VCore when overclocked at 2.7-3.0GHz
You can't compare cpus of different architectures, or even cpus of the same architecture, but on a different process. 90nm K8 maxed out at around 1.55V safely, 65nm K8 was only good to about 1.5V safely. These are modified k8's built on 45nm. I wouldn't go above 1.45V for 24/7, personally.
Posted on Reply
#97
cdawall
where the hell are my stars
Wile EBecause most people don't OC. 140w is most definitely an issue when it comes to OEMs.



You can't compare cpus of different architectures, or even cpus of the same architecture, but on a different process. 90nm K8 maxed out at around 1.55V safely, 65nm K8 was only good to about 1.5V safely. These are modified k8's built on 45nm. I wouldn't go above 1.45V for 24/7, personally.
OEM's will never use this chip.

outside of falcon northwest, alienware and voodoo who uses black edition or extreme edition cpu's? this chip will never be in the mainstream market. the companies that will use these chips will put them in higher end boards such as alienware who will use an asus M4A79 series board.

and AMD has officially spec'd K10 45nm to run 1.5v stock and unofficially announced 1.55v is safe on good air cooling

just check newegg under the phenom 955

Posted on Reply
#98
ShadowFold
I run my 720 at 1.52v 24/7. Stays nice and cool :)
Posted on Reply
#99
Wile E
Power User
ShadowFoldI run my 720 at 1.52v 24/7. Stays nice and cool :)
Heat has absolutely nothing to do with it. It has to do with electron migration. If you are past a "safe" 24/7 voltage, even sub-zero temps won't stop degradation.

But if AMD says 1.5V, who am I to argue?
Posted on Reply
#100
cdawall
where the hell are my stars
Wile EHeat has absolutely nothing to do with it. It has to do with electron migration. If you are past a "safe" 24/7 voltage, even sub-zero temps won't stop degradation.

But if AMD says 1.5V, who am I to argue?
lol no kidding if they say 1.5v is truly safe 1.55v should be ok for about 5 years
Posted on Reply
Add your own comment
Nov 24th, 2024 07:30 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts