Intel to reduce P4 to 65w

[$BangforThe$]

Intels Strained Silicon seems to be very health on the 65nm shrink .

http://www.xbitlabs.com/news/cpu/display/20060421031743.html

 

[Dillusion]

Good news for people still wanting to buy netburst architecture at those new insane prices, good to hear.

 

[mwarps]

Intels Strained Silicon seems to be very health on the 65nm shrink .

http://www.xbitlabs.com/news/cpu/display/20060421031743.html

Strained silicon has almost nothing to do with power consumption and everything to do with Transistor speed.

They must have stopped using Oxide as the gate dielectric and gone to something HiK. I am wondering how they overcame the scattering and the nasty band bending.

 

[Asian Dub Foundation]

if those price drops are enough, its not a bad buy

 

[MrGuvernment]

yeah, good by celeron :d or hello even cheaper celeron low end systems!

AMD just keeps hearing more and more bad news.

 

[drizzt81]

Good news for people still wanting to buy netburst architecture at those new insane prices, good to hear.

yeah, that may be breathing some more life into my old SC420. I'd love to drop in a 6xx series P4 and replace the old, old celery that's in it. Now I must pray for mobo compatibility

 

[Digital Viper-X-]

I Dunno if you guys have seen yet but an AM2 3500 + is about the same price as a 2.8single core p4 o.0

prices are going insane for single core cpus on both ends and intel already has alot of low priced dual core <3 my new 920D

 

[ZenOps]

I'm already halfway there with my Pentium 631.

All I did was undervolt from the stock voltage of 1.25volts, to 1.1875Volts which is the minimum the Asus P5LD2-VM will let me do. If the TDP is 86watts at 1.25 volts, at 1.1875 the TDP should be around 77.6 watts.

This allows me to use the stock intel heatsink and undervolt the fan to 7 volts (2100 RPM max) and its fairly quiet.

Still overclocks from 3.0 to 3.7Ghz with the reduced voltage too. I find the limiting factor is the quality of the Northbridge PCI-E which must scale up with the processor or else it becomes unstable.

Asus P5LD2-VM
Intel Pentium 631 undervolted to 1.1875 volts and overclocked to 3.7Ghz
1GB OCZ DDR2 at 1.814 volts
Evga 6800GS PCI-E running standard voltage.
SB Audigy 2zs
Seagate 7200.9
Benq LS/DVD 1655

Draws 170 watts at wall outlet when playing 3D games. Power supply is a 300 watt Fortron which at approximately 70 percent efficiency needs only provide 120 watts... I had this at one time running off an Antec Aria 300watter, but finally decided to move back up to a regular mid-ATX to keep things quieter.

 

[Terra]

I need to get something clarified about TDP.
Intel messuares TPD at max CPU load.
AMD messuares TPD at a lesser load...or am I mistaken?

Terracide...

 

[mwarps]

I need to get something clarified about TDP.
Intel messuares TPD at max CPU load.
AMD messuares TPD at a lesser load...or am I mistaken?

Terracide...

I don't think so. I am too busy at work to go find tech docs.
But my dual Opty 165 at 2.4 only sucks down 159-164W for the whole system depending on the day, and I believe the CPU is rated for 89W.

Kill-a-watt tested, me approved.

EDIT: at full load.

 

[Terra]

I don't think so. I am too busy at work to go find tech docs.
But my dual Opty 165 at 2.4 only sucks down 159-164W for the whole system depending on the day, and I believe the CPU is rated for 89W.

Kill-a-watt tested, me approved.

EDIT: at full load.

I am not talking about the Watt usage.
I am talking about the TDP

Terra...

 

[mwarps]

I am not talking about the Watt usage.
I am talking about the TDP

Terra...

Courtesy of FreiDog:

AMD defines TDP (at least on K8) as an absolute maximum rating (Idd_max * Vcc_nom) for an entire group of chips.

 

[visaris]

I need to get something clarified about TDP.
Intel messuares TPD at max CPU load.
AMD messuares TPD at a lesser load...or am I mistaken?

Intel's TDP describes the "average" TDP.
AMD's TDP describes the "worst case" TDP.

So, AMD 65w TDP will use less power than Intel 65w TDP.

AMD has said they have plans to change their TDP figure to "average" in the future as a means to remove Intel's TDP marketing advantage.

 

[Terra]

Intel's TDP describes the "average" TDP.
AMD's TDP describes the "worst case" TDP.

So, AMD 65w TDP will use less power than Intel 65w TDP.

AMD has said they have plans to change their TDP figure to "average" in the future as a means to remove Intel's TDP marketing advantage.

Ah...I got it backwards...thanks for the info

Terra...

 

[visaris]

How does AMD define TDP?

“Thermal Design Power (TDP) is measured under the conditions of TCASE Max, IDD Max, and VDD=VID_VDD, and include all power dissipated on-die from VDD, VDDIO, VLDT, VTT, and VDDA.”

This means that TDP, as defined by AMD, is measured at the maximum current the CPU can draw, at the default voltage, under the worst-case temperature conditions. This is the maximum power that the CPU can possibly dissipate. Intel, however, has a different definition.

How does Intel define TDP?

From the Intel Datasheet for Northwood CPUs:

“The numbers in this column reflect Intel’s recommended design point and are not indicative of the maximum power the processor can dissipate under worst case conditions.”

And from Intel’s datasheet for Prescott CPUs:

“Thermal Design Power (TDP) should be used for processor thermal solution design targets. The TDP is not the maximum power that the processor can dissipate.”

And the most telling quote of all, contained in both documents:

“Analysis indicates that real applications are unlikely to cause the processor to consume maximum power dissipation for sustained periods of time. Intel recommends that complete thermal solution designs target the Thermal Design Power (TDP) indicated in Table 26 instead of the maximum processor power consumption. The Thermal Monitor feature is intended to help protect the processor in the unlikely event that an application exceeds the TDP recommendation for a sustained period of time.”

What this means is that Intel’s TDP is actually lower than the maximum power dissipation of the processor. This is in stark contrast to AMD’s TDP numbers, which are higher than the respective processor’s maximum power dissipation.

 

[visaris]

Still, Intel selling a P4 at 65w is much more attractive. The number one problem with the P4 IMHO was heat. Good job Intel. If they could have done this a year earlier they would have made a lot more money.

 

[mwarps]

How does AMD define TDP?

“Thermal Design Power (TDP) is measured under the conditions of TCASE Max, IDD Max, and VDD=VID_VDD, and include all power dissipated on-die from VDD, VDDIO, VLDT, VTT, and VDDA.”

This means that TDP, as defined by AMD, is measured at the maximum current the CPU can draw, at the default voltage, under the worst-case temperature conditions. This is the maximum power that the CPU can possibly dissipate. Intel, however, has a different definition.

How does Intel define TDP?

From the Intel Datasheet for Northwood CPUs:

“The numbers in this column reflect Intel’s recommended design point and are not indicative of the maximum power the processor can dissipate under worst case conditions.”

And from Intel’s datasheet for Prescott CPUs:

“Thermal Design Power (TDP) should be used for processor thermal solution design targets. The TDP is not the maximum power that the processor can dissipate.”

And the most telling quote of all, contained in both documents:

“Analysis indicates that real applications are unlikely to cause the processor to consume maximum power dissipation for sustained periods of time. Intel recommends that complete thermal solution designs target the Thermal Design Power (TDP) indicated in Table 26 instead of the maximum processor power consumption. The Thermal Monitor feature is intended to help protect the processor in the unlikely event that an application exceeds the TDP recommendation for a sustained period of time.”

What this means is that Intel’s TDP is actually lower than the maximum power dissipation of the processor (and as you’ll see later, it can be significantly lower). This is in stark contrast to AMD’s TDP numbers, which are higher than the respective processor’s maximum power dissipation.

So unless they have changed this methodology, the Conroe may not be as "power sipping" as I previously thought.

So for the conditions I run at,

65W Intel > 65W AMD.

Thanks, visaris. Good to know.

 

[visaris]

Ah...I got it backwards...thanks for the info

Thanks, visaris. Good to know.

No problem; I'm glad I could help.

 

[ZenOps]

IMO, a quick general way to tell how hot a processor is going to get is by the die size and the voltage (and to a small extent the mhz frequency)

Dies with 4MB+ (in a 2x2 arrangement like the 9xx or a sharable like in Conroe) cache are quite large, and will put out heat just based on die size and voltage.

That the newer intel 6xx, 9xx, and Conroes have the ability to speedstep probably simply means that if they ever go over the TDP they will simply undervolt and either underclock and/or clock throttle to compensate.

I do want a newer 631 though, with the *not broken* speedstep... I never use dual core in gaming, so duallies is not even an issue with me (no wonder both Intel and AMD are dropping prices on duallies)

 

[coldpower27]

IMO, a quick general way to tell how hot a processor is going to get is by the die size and the voltage (and to a small extent the mhz frequency)

Dies with 4MB+ (in a 2x2 arrangement like the 9xx or a sharable like in Conroe) cache are quite large, and will put out heat just based on die size and voltage.

That the newer intel 6xx, 9xx, and Conroes have the ability to speedstep probably simply means that if they ever go over the TDP they will simply undervolt and either underclock and/or clock throttle to compensate.

I do want a newer 631 though, with the *not broken* speedstep... I never use dual core in gaming, so duallies is not even an issue with me (no wonder both Intel and AMD are dropping prices on duallies)

That actually doesn't make much sense, caches are typically the lowest power draining component of the CPU, it's the processor core logic that is draining all the energy.

Conroe actually puts out relatively little heat, and since you have to factor in Conroe and Presler are on 65nm processes they have smaller die sizes then AMD's Dual Cores. Which have smaller caches.

 

[coldpower27]

So unless they have changed this methodology, the Conroe may not be as "power sipping" as I previously thought.

So for the conditions I run at,

65W Intel > 65W AMD.

Thanks, visaris. Good to know.

It varies from processor to processor, Intel's use of TDP is not exactly constant. I would have to say this though for instance Presler XE and Windsor FX's TDP's of 125W and 130W are actually pretty accurate with one another.

While if you compare a Presler XE 965 130W to a Smithfield XE 840 130W, they are vastly different in energy consumption.

I will say this though the Conroe XE's 75W doesn't use the same scale as Windsor FX 125W TDP.

 

[pxc]

So unless they have changed this methodology, the Conroe may not be as "power sipping" as I previously thought.

From a more recent Intel data sheet (I found this in 2005, but I don't remember how new it was):

The processor’s power is specified as Thermal Design Power (TDP) for thermal solution design. TDP is defined as the worst-case power dissipated by the processor while executing publicly available software under normal operating conditions, at nominal voltages that meet the load line specifications.

Parsed:
"worst-case power dissipated by the processor while executing publicly available software" (not a specially crafted "power virus"... dual Prime 95 or any other software you run is not a "power virus". load up the cores at 100% with any software you can find and it's still not a "power virus")
"normal operating conditions" (not throttling Tcase < Tcasemax)
"nominal voltage" (standard voltage within tolerances)
"meet the load line specifications" (sufficient current)

So for the conditions I run at,

65W Intel > 65W AMD.

That's true sometimes, but not really due to the "conditions." A top speed 65W TDP Intel CPU, which is actually a platform rating (see above)*, will likely use more power than a top speed 65W A64. It can vary for lower speed chips. Unless your conditions happen to mean running (edit) an energy efficient 65W AM2 4600+ vs Conroe 2.66GHz at stock speeds with both cores at 100% load, I wouldn't rush to make any big judgements.

* lower speed chips will generally use less power than higher speed chips. i.e. 1.86GHz Conroe is "65W TDP" and so is the 2.66GHz Conroe. Intel and AMD can of course sell leaky/higher voltage chips at lower speed grades instead of scrapping parts, making some lower speed chips use the same power as faster chips.

Strained silicon has almost nothing to do with power consumption and everything to do with Transistor speed.

http://developer.intel.com/technology/silicon/power/transistor.htm

Strained Silicon Helps Control Transistor Leakage

Reduced leakage current means better power efficiency and less heat dissipation per transistor.
...
Intel's proprietary "uni-axial" strained silicon reduced leakage current by five times or more without diminishing performance (for example, on-current).

Leakage is a big part of power consumption, which is why AMD went SOI and Intel went strained silicon.

 

[$BangforThe$]

Thanks pxc . I was so stunned by the high K remark I just couldn't reply.

 

[drizzt81]

from:
Intel® Pentium® D Processor 950, 940, 930, 920&#916;, Intel® Pentium® D Processor 840, 830 and 820&#916; and Intel® Pentium® Processor Extreme Edition 840&#916; Thermal and Mechanical Design Guidelines (TMDG) (April 2006) page 12

Thermal Design Power (TDP) is a power dissipation target based on worst-case
applications (my emphasis). Thermal solutions should be designed to dissipate the thermal design
power.

Interpret that whichever way you want. Someone else do the AMD research, I have to do some work.

 

[Terra]

from:
Intel® Pentium® D Processor 950, 940, 930, 920&#916;, Intel® Pentium® D Processor 840, 830 and 820&#916; and Intel® Pentium® Processor Extreme Edition 840&#916; Thermal and Mechanical Design Guidelines (TMDG) (April 2006) page 12


Interpret that whichever way you want. Someone else do the AMD research, I have to do some work.

So I was right about Intel's TDP?

Terra - Confusing thread *L*

 

[drizzt81]

So I was right about Intel's TDP?

Terra - Confusing thread *L*

Well, it depends how you want to interpret the meaning of "target" in my previous quote:

• Our target is 95W, what we do in the real world, is a different story
• Our target is to stay below 95W; it is conceivable that someone writes a virus or something else that will put out >95W, so we had to put the word target in here to make the lawyers happy. It's not going to happen in 99.9999% of the cases

bottom line: we will never know, since Intel doesn't want to be held liable should someone find a way to have a proc dissipate > 95W.

However, let's consider Intels motivation in the wording: On the one hand they'd like a low number, since it looks cool for the people buying it. On the other hand, if their number is too low, the OEMs and System Integrators would design insufficient cooling, which will negatively effect Intels image: systems overheat and crash, which reflects poorly on Intel's "super-stable platform" image.

Considering how many more people would buy a P4 if it had a lower TDP vs. the damage to Intels image if their systems go down the drain, I would say that the TDP figure is likely very close to the actual max that will be dissipated.

I am open for different oppinions, but I would like to either see some factual proof or at least decent reasoning.

 

[Terra]

Well, it depends how you want to interpret the meaning of "target" in my previous quote:

• Our target is 95W, what we do in the real world, is a different story
• Our target is to stay below 95W; it is conceivable that someone writes a virus or something else that will put out >95W, so we had to put the word target in here to make the lawyers happy. It's not going to happen in 99.9999% of the cases

bottom line: we will never know, since Intel doesn't want to be held liable should someone find a way to have a proc dissipate > 95W.

However, let's consider Intels motivation in the wording: On the one hand they'd like a low number, since it looks cool for the people buying it. On the other hand, if their number is too low, the OEMs and System Integrators would design insufficient cooling, which will negatively effect Intels image: systems overheat and crash, which reflects poorly on Intel's "super-stable platform" image.

Considering how many more people would buy a P4 if it had a lower TDP vs. the damage to Intels image if their systems go down the drain, I would say that the TDP figure is likely very close to the actual max that will be dissipated.

I am open for different oppinions, but I would like to either see some factual proof or at least decent reasoning.

So I was right in 99.9999% of the cases?

Terra - Know I know even less than when I first asked *L*

 

[pxc]

Someone else do the AMD research, I have to do some work.

AMD has specified it two main ways:

Thermal Design Power represents the maximum sustained power dissipated while executing publically-available software or instruction sequences under normal system operation at nominal Vcc_core. Thermal solutions must monitor the temperature of the processor to prevent the processor from exceeding its maximum die temperature.

and

4. Thermal Design Power (TDP) is measured under the conditions of Tcase Max, IDD Max, and VDD=VID_VDD, and include all power dissipated on-die from VDD, VDDIO, VLDT, VTT, and VDDA.
...
8. Thermal Design Power dissipated by the processor VDDIO, VTT, VLDT, and VDDA power planes only.
9. Thermal Design Power dissipated by the processor VDDIO and VTT power planes only.
...
14 Thermal Design Power (TDP) and IDD max for Rev E and later processors are the limits at the highest Tcase max in the specified range for the corresponding OPN. Products will conform to the TDP and IDD Max limits at all valid voltages. The relationship of Tcase max and Thermal Profile to TDP for a specific device is defined in Notes:
Table 8.

The current data sheet doesn't really say what IDD Max represents (instantaneous current or a sustained current on full processor core utilization). Older Athlon 64 data sheets had the first quote above which is close to Intel's definition of TDP. Notice that IDD Max * VID_VDD is not equal to the TDP rating on AMD's data sheets. Like Intel, AMD has a margin of safety (and consistency) built into a TDP platform rating.

It's a bit silly to call IDD_max * VID_VDD the "real power" anyways. IDD_max is an instantaneous maximum current on Intel's data sheets, not continuous. That value is a concern to the manufacturer of a motherboard (those caps around the CPU and multiple power traces/widths aren't for show), not for cooling. Some people think TDP is a scam, but it's actually very meaningful. It's made to answer an important question: how much cooling do I need use to run a chip in the worst possible case without resorting to throttling?
----

drizzt81: A "power virus" is not really a virus. I actually have a hard time imagining how such a power virus could run under a modern OS without crashing it (not due to heat, but due to suspending all other processing including the OS). And even then, both AMD and Intel processors throttle when the core temperature gets too high. If one did exist it's unlikely to cause any damage other than locking up the PC until it's reset. The same effect can be done in much easier ways. And finally, many LGA775 cooling solutions are made to handle up to (or over) 130W TDP chips since Intel has released some with that rating. I wouldn't worry about a 95W TDP chip being in any danger.

 

[Donnie27]

AMD has specified it two main ways:

and
The current data sheet doesn't really say what IDD Max represents (instantaneous current or a sustained current on full processor core utilization). Older Athlon 64 data sheets had the first quote above which is close to Intel's definition of TDP. Notice that IDD Max * VID_VDD is not equal to the TDP rating on AMD's data sheets. Like Intel, AMD has a margin of safety (and consistency) built into a TDP platform rating.

It's a bit silly to call IDD_max * VID_VDD the "real power" anyways. IDD_max is an instantaneous maximum current on Intel's data sheets, not continuous. That value is a concern to the manufacturer of a motherboard (those caps around the CPU and multiple power traces/widths aren't for show), not for cooling. Some people think TDP is a scam, but it's actually very meaningful. It's made to answer an important question: how much cooling do I need use to run a chip in the worst possible case without resorting to throttling?
----

drizzt81: A "power virus" is not really a virus. I actually have a hard time imagining how such a power virus could run under a modern OS without crashing it (not due to heat, but due to suspending all other processing including the OS). And even then, both AMD and Intel processors throttle when the core temperature gets too high. If one did exist it's unlikely to cause any damage other than locking up the PC until it's reset. The same effect can be done in much easier ways. And finally, many LGA775 cooling solutions are made to handle up to (or over) 130W TDP chips since Intel has released some with that rating. I wouldn't worry about a 95W TDP chip being in any danger.

I looked in vain for the story about during the Fall of 2005 that Intel was moving to an industry standard that they already use on other hardware.

Folks on AMD's side have always used that Intel = Normal/Average Temps and AMD = Max in worst case, that's nothing new. The way to solve that is always is look at the end result.

HKEPC

AMD Athlon 64 FX 62 (AM2) 2.8GHz/1MB x2 TDP 125W, Idle 98W 142W Max temp measured.

Intel Core 2 Extreme X6800 2.93GHz/4MB Intel Windows XP TDP 75W, Idle 77W 118W Max "

Pentium XE 965 3.73GHz/2MB x2 CPU Burn x2 130W TDP, Idle 118W 147W Max "

So IF AMD says their 3GHz will have a TDP of 80W, thats max it would look like TDP 80W, about 65W IDLE and about 100W (est) load worst case. 80 watts is just an example. The difference from idle to load is a 41W increase for Intel and a 44 W increase for AMD above. BOTH went past the TDP.

Core 2 Duo extreme 35.5C idle and 49.2C under load.
A65 FX 62 42.2C idle and 57.4C under load

So the TDP arguement is dead until AMD reaches 65nm.

 

[drizzt81]

----

drizzt81: A "power virus" is not really a virus. I actually have a hard time imagining how such a power virus could run under a modern OS without crashing it (not due to heat, but due to suspending all other processing including the OS). And even then, both AMD and Intel processors throttle when the core temperature gets too high. If one did exist it's unlikely to cause any damage other than locking up the PC until it's reset. The same effect can be done in much easier ways. And finally, many LGA775 cooling solutions are made to handle up to (or over) 130W TDP chips since Intel has released some with that rating. I wouldn't worry about a 95W TDP chip being in any danger.

Virtually independent of whether it's possible or not, it's clear that in todays 'sue-happy' society and company will make sure that their performance claims etc. are vague enough to get out of a lawsuit. I am not concerned about hte TDP at all, since the only non-mobile Intel chip I currently own is a Celeron 3xx and that came with my dell... I just wanted to provide a link for people to "hear it from the horse's mouth".
That way there cannot be any argument about the source being biased or wrong. If Intel's definition of TDP is what I quoted above, then that is what Intel means when saying "The TDP of this processor is 65W".
People can argue about what excatly the statement "means" until they are blue in the face, if they so desire; At least they have a fact that they can try and bend whichever way they want and don't argue over some information that they have had their 3rd cousin's nephew tell them once...

 

[pxc]

They way to solve that as always is look at the end result.

Yeah, that's what matters. It's just that I hear so much bad information about TDP it gets ridiculous and I post.

I'm glad you found an example of 2 high end chips. I don't know if they have the measurement right (the meter can only measure the output... how and where it's used on the board can vary and the FX62 and Core 2 XE values make little sense for idle, plus it also includes the losses for the voltage regulators on the board), it's still good enough for relative comparison between the FX62 and 965XE. It points out what I said above, that you can compare the highest models at a given TDP. Lower speed chips, even if they have the same TDP as the highest speed models, will on average use less power.
----

Back to the original topic... I wish 965 boards were out already. I'd buy one and a cheap 6x1 processor to hold me off until Conroe, and recoup a bit from my X2 + motherboard (both were cheap... $240 @ dell + $45 @ monarch).

 

[coldpower27]

I looked in vain for the story about during the Fall of 2005 that Intel was moving to an industry standard that they already use on other hardware.

Folks on AMD's side have always used that Intel = Normal/Average Temps and AMD = Max in worst case, that's nothing new. The way to solve that is always is look at the end result.

HKEPC

AMD Athlon 64 FX 62 (AM2) 2.8GHz/1MB x2 TDP 125W, Idle 98W 142W Max temp measured.

Intel Core 2 Extreme X6800 2.93GHz/4MB Intel Windows XP TDP 75W, Idle 77W 118W Max "

Pentium XE 965 3.73GHz/2MB x2 CPU Burn x2 130W TDP, Idle 118W 147W Max "

So IF AMD says their 3GHz will have a TDP of 80W, thats max it would look like TDP 80W, about 65W IDLE and about 100W (est) load worst case. 80 watts is just an example. The difference from idle to load is a 41W increase for Intel and a 44 W increase for AMD above. BOTH went past the TDP.

Core 2 Duo extreme 35.5C idle and 49.2C under load.
A65 FX 62 42.2C idle and 57.4C under load

So the TDP arguement is dead until AMD reaches 65nm.

Well for now let's assume that it is unchanged and the Athlon FX 64 should have a 125W TDP like the FX 62 Windsor does now. The differnece here between Conroe XE 75W TDP and Windsor is 50W but the load difference is translating to 24W.

Though like I have said before, The current Windsor FX 62 125W TDP and Presler XE 965 130W TDP match pretty damn well to each other.

And no they haven't breached their TDP necessarily, those are system power level draws. Not CPU on it's own power level draws, we don't know exactly how much Conroe XE or Windsor FX draw on their own right now, since it's very difficult to just isolate just the CPU by itself.

 

[coldpower27]

So I was right in 99.9999% of the cases?

Terra - Know I know even less than when I first asked *L*

I would say TDP's are jsut general guidelines as a place to start to look for power conusmption numbers though you have to rely on measure numbers to see what the actual draw is.

 

[Terra]

I would say TDP's are jsut general guidelines as a place to start to look for power conusmption numbers though you have to rely on measure numbers to see what the actual draw is.

I belived that the TDP has to do with heat disapation...not power consumption?

Terra...

 

[coldpower27]

I belived that the TDP has to do with heat disapation...not power consumption?

Terra...

Power conusmption and heat dissappation are related items, so if a chip is spewing out more heat you can expect it to cosume more power. Hence a generall ballpark guideline of where to start looking.

 

[pxc]

Not CPU on it's own power level draws, we don't know exactly how much Conroe XE or Windsor FX draw on their own right now, since it's very difficult to just isolate just the CPU by itself.

Too bad motherboard manufacturers don't allow monitoring directly. That's pretty much the only place where it can be isolated.

I don't put much into those HKPC numbers because the PSU is only monitoring a 12v rail output. Under load using the same cooler (as a cross check), the X6800 is running 8.2C cooler than the FX62. That implies a ~31W power difference between the X6800 and FX62, not even counting VR losses or the less effective transfer of heat due to the smaller die size which would make the difference even higher. Those results are just weird.

 

[coldpower27]

Too bad motherboard manufacturers don't allow monitoring directly. That's pretty much the only place where it can be isolated.

I don't put much into those HKPC numbers because the PSU is only monitoring a 12v rail output. Under load using the same cooler (as a cross check), the X6800 is running 8.2C cooler than the FX62. That implies a ~31W power difference between the X6800 and FX62, not even counting VR losses or the less effective transfer of heat due to the smaller die size which would make the difference even higher. Those results are just weird.

Hmm that is interesting, maybe they should do the same thing but monitoring how much power the system draws from the power outlet using 2 Seasonic S12 600W PSU to help minimize efficiency loses.

That is also assuming they mounted the heatsink properly on both processors. I don't prefer using temperature as that has the capacity to be affected but alot more things.

 

[mwarps]

*snip*
Judgement on conroe TDP laid until I can plug it into the wall.
*/snip*

http://developer.intel.com/technology/silicon/power/transistor.htm
Leakage is a big part of power consumption, which is why AMD went SOI and Intel went strained silicon.

Nice. Someone with half a brain. AMD is using SSOI. For some reason there's some off acronym I can't remember which they're using.

But if you read the article, it states that the High K dielectric has reduced the leakage 100x, as opposed to the 5x of the strained silicon.

Who's right? I should know, but I forgot which is more prevalent in smaller devices - gate leakage or s/d leakage.

 

[Donnie27]

Well for now let's assume that it is unchanged and the Athlon FX 64 should have a 125W TDP like the FX 62 Windsor does now. The differnece here between Conroe XE 75W TDP and Windsor is 50W but the load difference is translating to 24W.

Though like I have said before, The current Windsor FX 62 125W TDP and Presler XE 965 130W TDP match pretty damn well to each other.

And no they haven't breached their TDP necessarily, those are system power level draws. Not CPU on it's own power level draws, we don't know exactly how much Conroe XE or Windsor FX draw on their own right now, since it's very difficult to just isolate just the CPU by itself.

I know and my bad about the system frackin opps!

 

[$BangforThe$]

Nice. Someone with half a brain. AMD is using SSOI. For some reason there's some off acronym I can't remember which they're using.

But if you read the article, it states that the High K dielectric has reduced the leakage 100x, as opposed to the 5x of the strained silicon.

Who's right? I should know, but I forgot which is more prevalent in smaller devices - gate leakage or s/d leakage.

Why you keep talking about High K Intel isn't using high K yet.