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

[Dippyskoodlez]

== memory sub-system/bandwidth fail.

Exactly.

Disabling part of the CPU to remain within TDP is not a viable option with a high end part.

 

[1Kurgan1]

I don't really see the issue as most people are OC'ing who buy these chips as AMD's don't come in many built factory computers. Odds are most people are pushing this kind of wattage already.

 

[Mussels]

I don't really see the issue as most people are OC'ing who buy these chips as AMD's don't come in many built factory computers. Odds are most people are pushing this kind of wattage already.

overclockers and enthusiasts are probably 5% of the market. lots - and i mean lots - of people just buy the high end parts and leave them at stock, because its the fastest they can buy.

 

[BrooksyX]

Dang 140w, thats pretty high.

 

[eidairaman1]

I don't really see the issue as most people are OC'ing who buy these chips as AMD's don't come in many built factory computers. Odds are most people are pushing this kind of wattage already.

so what about your machine?

 

[mtosev]

AMD did it again.

the cpu has a higher TDP then an i7 cpu and also has more MHz but is still shower then an i7 cpu whitch has less MHz.

AMD FAILED.

 

[Darren]

mtosev,

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.

 

[eidairaman1]

im on a old machine myself and will be upgrading at the end of the year, the Core i7 seems to be out of my range and i don't need all that processing for what i do anyway.

BTW TheMailMan78 you have another plus 1 for your wise cracks

 

[Mussels]

darren: please dont double post, and try not to post inflammatory material.

If you respond to someone baiting you, it gets both people in trouble. If you ignore them, only they do.

 

[wiak]

Cmon now. It's nowhere near 200W tdp (I'm serious, intel lied so bad back then)

they still do
heck a intel 95W cpu is a AMD 125W cpu, its just how intel calulates *load*
btw the Phenom II 965 isnt even released yet, last thing i heard its a month away from release, and msi could have gotten a 125W part and forgot to update the cpu support list

 

[hat]

I wish they would just multiply voltage by whatever amperage the processor sucks up...

 

[Mussels]

AMD dont use real watts either, both companies use TDP.

I dont have a huge argument against it either -looking at my video cards, furmark uses 100W more than any game, no matter the settings i use, so if AMD wanted to label it with a TDP of less than its max, that makes sense to me.

Same applies with CPU's, the odds on 100% power usage outside of stress testing is pretty much nil.

 

[Deleted member 24505]

How exactly is TDP calculated?

 

[mtosev]

How exactly is TDP calculated?

i think its based on estimates.

 

[FordGT90Concept]

phenom II is a power efficient architecture, instead intel's core i7 is a certified powerhog, temps under load are 80 c for core i7 with stock cooler and stock speed , while phenom II is just 45 c under load with stock cooler and stock speed. everybody that have core i7 has to suffer the heat and the price (like in hell) while phenom II users are cool and with money in the wallet.

I think Core i7 is more efficient for several reasons:
-Core i7 965 (130w TDP) is lower power than Phenom 965 (140w TDP).
-Core i7 965 (~1.1v) does more work at fewer volts than Phenom II 965 (~1.2v).
-Core i7 965 (80-90C) runs hot with HT enabled because more of the chip is being used (signifying architecture efficiency).
-Phenom 965 can't hold a candle to the Core i7 965 in terms of performance (only exception being high resolution gaming).

How exactly is TDP calculated?

It is more like a specification (Thermal Design Power). AMD figures out the most this processor could safely draw and that establishes the TDP. From there, they decide how heavy of an HSF is needed to dissipate as much heat as the TDP suggests and also, motherboards have to have enough voltage regulators to handle that high of a load. TDP is determined by AMD/Intel to signify to the rest of the industry what it will take to run and cool that processor (or chipset).

 

[Mussels]

How exactly is TDP calculated?

no one knows. They say its an 'average' load number.

 

[TheLaughingMan]

Hey

http://img.techpowerup.org/090710/Capture027286.jpg

100$ cheaper and same overclocking performance.

Wow, that build is my computer but, I got 1333 RAM because it was on sale from G.Skills for only $55. I will be trying to OC it to 1600 and see how it holds up.

 

[Dippyskoodlez]

no one knows. They say its an 'average' load number.

TDP is supposed to be worst case scenario. Hence why we see much lower numbers on average.

Intel and AMD also calculate their TDP differently, so they are not comparable directly.

 

[TheMailMan78]

TDP is supposed to be worst case scenario. Hence why we see much lower numbers on average.

Intel and AMD also calculate their TDP differently, so they are not comparable directly.

Unless you turn "Cool and Quiet" off. Then it stay at max all the time.

 

[Dippyskoodlez]

Unless you turn "Cool and Quiet" off. Then it stay at max all the time.

Max should not be "worst case scenario", not all CPU's are exactly the same.

i.e. TWKR.

 

[TheMailMan78]

Max should not be "worst case scenario", not all CPU's are exactly the same.

Well of course but every chip is rated in a certain power spectrum. This is why the chip we are talking about is 140w. Of course I could be wrong. I'm a noob making educated guesses

 

[Flyordie]

Exactly.

Disabling part of the CPU to remain within TDP is not a viable option with a high end part.

Making a pure AM3 part wouldn't hurt...
Disabling the DDR2 controller would save 10-15W as seen in the comparison of the same CPU on the 2 different platforms.

 

[Dippyskoodlez]

Well of course but every chip is rated in a certain power spectrum. This is why the chip we are talking about is 140w. Of course I could be wrong. I'm a noob making educated guesses

Well you have to consider AXP's were always rated the same.

An AXP 2600+ was rated the same TDP was the 3200+.

There's a baseline margin of error for an absolute piss poor quality CPU that is added for a safety margin and to reduce the number of deaths/DOA chips. Then there is the real performance of chips which will change vastly too. TWKR chips were reported as "high leakage" parts. TDP only refers to the absolute maximum a given grade of CPU should -ever- put out, and if designed right should never actually hit that threshold. (The exact method this is calculated is not given by AMD -OR- Intel. )

Desktop CPU's are often 85W chips.. but hardly any have ever actually held 85W of heat output before overclocking. Otherwise I should be able to use my AXP cooler on my A64.

AMD's TDP does give a good indicator of struggling with power consumption for CPU production.

 

[TheMailMan78]

Well you have to consider AXP's were always rated the same.

An AXP 2600+ was rated the same TDP was the 3200+.

There's a baseline margin of error for an absolute piss poor quality CPU that is added for a safety margin and to reduce the number of deaths/DOA chips. Then there is the real performance of chips which will change vastly too. TWKR chips were reported as "high leakage" parts. TDP only refers to the absolute maximum a given grade of CPU should -ever- put out, and if designed right should never actually hit that threshold. (The exact method this is calculated is not given by AMD -OR- Intel. )

Desktop CPU's are often 85W chips.. but hardly any have ever actually held 85W of heat output before overclocking. Otherwise I should be able to use my AXP cooler on my A64.

AMD's TDP does give a good indicator of struggling with power consumption for CPU production.

Thanks for the education.

 

[btarunr]

Making a pure AM3 part wouldn't hurt...
Disabling the DDR2 controller would save 10-15W as seen in the comparison of the same CPU on the 2 different platforms.

Uh 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.