DailyTech - "Merom" TDP Not All That

Submit News

Blog: Hardware "Merom" TDP Not All That
Sven Olsen (Blog) - August 17, 2006 2:18 PM

22 comment(s) - last by hstewarth.. on Aug 20 at 11:22 AM

"Merom" runs hot, "Yonah" might be the better chip for notebooks

Intel began releasing specifications of the new Merom processor to its motherboard partners today. For those expecting Merom to increase performance over Yonah while simultaneously decreasing the thermal envelope, think again.

The top of the line Yonah processor, the T2700, has a TDP of 31W at 2.33GHz clock frequency. All Core 2 Duo Merom processors have a TDP of 34W, including the lowly 1.66GHz T5500. In comparison, the 1.66GHz Yonah rings in at 27W.

Surely with Enhanced Speed Step these numbers get better, we'd think. In "Battery Mode," all Merom processors clock down to 1GHz and 0.75V -- yet amazingly the TDP is still 20W. Yonah, which also clocks down to 1GHz in Battery Mode with a 0.95V core has a TDP of 13.1W.

Those expecting to pop in a Core 2 Duo Merom processor to alleviate an overheating MacBook Pro, look not here. Merom is a better performing processor than Yonah, but its thermals on paper show its advantages are only in performance and not in thermals at all.

Intel will counter these poor thermals with more "Low Voltage" and "Ultra Low Voltage" processors. Just a few weeks ago, Intel announced an Ultra-Low Voltage (ULV) Yonah processor running on a TDP of just 9W. Intel's newest roadmap includes the U7500, an ultra-low voltage version of Merom. However, given the fact that the normal voltage Yonahs have a lower TDP than the average Merom processor, we'd be hard pressed to think U7500 could possibly run cooler than its Yonah predecessor.

Comments

Threshold

Username
Password
remember me

This article is over a month old, voting and posting comments is disabled

Ultra Low Voltage

By rqle on 8/17/2006 4:41:41 PM , Rating: 2

the ultra low voltage version suppose to be able to run most normal task and apps at under 1watt. and max TDP at 9w so i heard.

RE: Ultra Low Voltage

By Master Kenobi

(blog) on 8/17/06, Rating: 0

RE: Ultra Low Voltage

By PT2006 on 8/17/2006 4:46:42 PM , Rating: 2

Both power consumption and TDP are written in Watts.

Parent

RE: Ultra Low Voltage

By s12033722 on 8/17/2006 5:58:45 PM , Rating: 2

It seems likely that if the TDP goes up from one processor to the other, the average consumed power will also go up. It's certainly not required that this be the case, but it does seem likely.

Parent

RE: Ultra Low Voltage

By Hare on 8/18/2006 3:09:27 AM , Rating: 2

quote:
The writer is confusing TDP with power usage (wattage).

How are these two different? If your CPU takes 25W (power usage) it definately will have 25W TDP. The current and voltage you feed in to the cpu is the power/wattage and it is 99,99% converted to heat -> TDP. Energy does not disappear...

Parent

RE: Ultra Low Voltage

By sleeprae on 8/18/2006 7:04:44 PM , Rating: 2

Exactly. For processing units, the overwhelming majority of consumed power is converted to heat. Therefore, you can state with some certainty that power consumption is approximately equal to thermal dissipation.

Parent

RE: Ultra Low Voltage

By Strunf on 8/20/2006 9:28:18 AM , Rating: 2

No the TDP is an informative value for system builders so they know what their coolers have to withstand, Intel and AMD probably pick like 1000 or soo higher clocked CPUs from a family and then measure the power consumption under different power usage conditions (idle, used at 100% ...) add to it like 30% to be on the safe side and then say its x TDP for whole family.
The power consumption on the other hand is a measurable unit that depends on the frequency of a CPU, usage and so on. It’s different for all the CPUs and even 2 2.6GHz CPU may have different power consumptions under the same conditions while having the same TDP.

A CPU may have a TDP of over 100 W and never reach it even when used at 100%, this is truer on the lower clocked CPU since they get the same TDP has their higher clocked brothers.

Parent

Not a good sign

By psychobriggsy on 8/17/2006 7:59:51 PM , Rating: 2

If Intel could guarantee that all the Merom processors it could make were in lower TDPs at each power state, then they would have set a lower power state.

The fact that at 1GHz Merom could potentially consume 20W, compared to 13.1W for Yonah and 10W for Turion X2 (@800MHz, I think someone is being rather optimistic with the figures at AMD, I got the figures from http://www.laptoplogic.com/resources/detail.php?id... ) is not good information for a processor aimed at laptops.

We all know that when it comes to laptop processors, Intel's TDPs are possibly even more pessimistic than AMDs, i.e., they aren't reached very often. However an increase is not a good sign. However the extra processing capability might compensate.

Here's hoping that Intel don't repeat that slow creep towards hot laptop processors again over the next few years.

RE: Not a good sign

By KristopherKubicki

(blog) on 8/17/2006 8:10:03 PM , Rating: 2

Keep in mind that the 20W in your example is the thermal envelope, not the power consumption. But yes, all of your other points are essentially valid.

Parent

RE: Not a good sign

By Hare on 8/18/2006 3:14:27 AM , Rating: 2

People also need to realise that when you compare the two it should not be done mhz to mhz, rather compare the actual performance. The clockspeed is meaningless.

Even if merom at 2Ghz consumes more energy than a 2Ghz Yonah it still has a lot better performance per Watt since it is capable of doing a lot more work in that time -> work done, drops more quickly to low power mode. The max. figures don't tell the whole story.

Parent

RE: Not a good sign

By Hare on 8/18/2006 4:53:03 AM , Rating: 2

Example: 2.16Ghz Yonah vs 2.0Ghz Merom. Merom was +23.3% faster in 3dmark03 cpu-test. An app(s) that stresses both cores makes the difference even bigger...

Parent

Look at AT's comparison

By IntelUser2000 on 8/18/2006 1:14:49 AM , Rating: 2

You can see Core 2 Duo consumes less power overall as a laptop compared to Core Duo, at worst cases equal.

RE: Look at AT's comparison

By dagamer34 on 8/18/2006 1:25:13 PM , Rating: 2

Yeah, Merom just finishes tasks faster, it doesn't use less power while doing it though.

Parent

RE: Look at AT's comparison

By sleeprae on 8/18/2006 7:02:04 PM , Rating: 2

One could argue, however, that Merom is able to spend a larger percentage of its time in a low-power state as it is able to complete tasks more quickly. I'm not sure how much impact this would have in practice, though.

Parent

RE: Look at AT's comparison

By The Cheeba on 8/19/2006 10:50:23 AM , Rating: 2

I noticed thermals were lacking in the AT review? Or did I miss something?

Parent

Not again...

By Phynaz on 8/17/06, Rating: 0

RE: Not again...

By PT2006 on 8/17/2006 3:05:53 PM , Rating: 2

OK. But why have a larger TDP for a processor thats supposedly more efficient. If you design a processor with a TDP of 30W, all that means is you can expect it to have a maximum dissapation of 30W. Therefore, I am expecting Merom to dissapate 34W or whatever it says.

Parent

RE: Not again...

By KristopherKubicki

(blog) on 8/17/2006 3:07:41 PM , Rating: 2

quote:
TDP does not state how much power a CPU dissipates.
quote:

TDP is the measurement of how much heat a CPU can dissapate at a maximum. The power consumption of these CPUs is less than 1.1W, at least according to Intel's roadmap.

Parent

RE: Not again...

By MercenaryForHire on 8/18/2006 10:01:28 AM , Rating: 1

quote:
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. The TDP definition is synonymous with the Thermal Design Power (typical) specification referred to in previous Intel datasheets. The Intel TDP specification is a recommended design point and is not representative of the absolute maximum power the processor may dissipate under worst case conditions.

Parent

Anandtech says otherwise

By Strunf on 8/20/2006 9:00:45 AM , Rating: 2

Sorry but you’re comparing apple to oranges, like it was said before the CPU are quite different in terms of performance per watt, hence if you want to speak of the best notebook CPU you have to take the performance of each of them into account.
The fact you say “All Core 2 Duo Merom processors have a TDP of 34W, including the lowly 1.66GHz T5500.”, should be a good reason to not come up with conclusion based on the TDP only, this cause the 1.66GHz T5500 probably wont use as much power as the 2.33GHz T7600.

The Anandtech article about the two also states on the conclusion “Overall, Merom may not be as big of an upgrade to Yonah as Conroe was to NetBurst, but the bottom line is that you get equal or better performance in every test without increasing cost or decreasing battery life.”

RE: Anandtech says otherwise

By hstewarth on 8/20/2006 11:22:58 AM , Rating: 2

Also the Merom can doe stuff that the Core Duo can't do. For example with Lightwave 3d 9.0, I can't run Core Duo Notebook in 64 bit more - but I can with Merom. Yes the uses for 64 bit are limited now - but it is something that the Core Duo can't do.

Also the Merom is based on Core 2 technology and Core 2 technology SSE instructions can execute in 1 clock cycle. So applications that use it will be much faster.

Parent

Upgrade macbook?

By Doormat on 8/17/2006 4:20:04 PM , Rating: 2

I thought the chips were soldered to the mobo? Upgrading the Mac Mini is more like it.

"If you look at the last five years, if you look at what major innovations have occurred in computing technology, every single one of them came from AMD. Not a single innovation came from Intel." -- AMD CEO Hector Ruiz in 2007