Power consumption question

There was a mention a while back that Core Duo would get about 30-60 minutes more power on a 8 cell battery versus an AMD x2. I can't find who originally posted it. And most articles I read state Intel is usually better with power consumption.



I just read this article on Core 2 Duo and it states power consumption isn't as great as Core Duo
http://www.dailytech.com/article.aspx?newsid=3840


If I'm considering a x2 TL-52 that's 31W ( according to
http://www.amd.com/gb-uk/Processors/ProductInformation/0,,30_118_13909_13911,00.html )

and the above daily tech article states all Merom processors are 34W,
how is it possible that battery life with a Core 2 Duo processor is significantly better than a AMD X2 ?

 

Well, that is people usually short cut their thinking. 65nm definitly consumes less power than 90nm A LOT. AMD gives us the exception that good architecture can beat good manufact some times.

Until Core 2 Duo releases to market and being fully reviewed, no body will really sure how this new design can buy you anything. Yeah, the dailytech article is a bit of scary, 34W Max TDP, and 20W in idle mode. This close to my undervolted AMD desktop TDP. I guess Intel just want to use this release to pickup the Gamming market, and use LV and ULV to win the Thin & Light market. LV? or ULV? how expensive would that be?

 

I was the original poster that said Core Duos would get 15-30 minutes more per 4 cell . There are two main reasons why Core Duo notebooks, even with a higher TDP, get more battery life than Turion X2 notebooks. The first reason is that TDP only refers to the processor's power consumption. While AMD's solution is better in this regard, remember that a notebook also has other components such as the the wireless card. Intel has greatly extended battery life by designing all these components in house and optimizing them for notebooks. AMD, on the other hand, does not make its own components besides processors and must rely on third parties to provide them. As in most cases, integration of parts by one company usually leads to better results. That is probably a reason why AMD recently bought ATI; working together, they can provide a complete package just like Intel's Centrino.

The second reason is that Core Duo is reasonably faster than the Turion X2 and thus finishes tasks faster. Even though the AMD processor looks better on paper, the Intel processor can revert back to a low power idle state faster than the Turion X2 and save battery life. The same concept also applies to Core Duos and Core 2 Duos in that Core 2 Duos may be more efficient and actually save battery life even with disadvantageous power specifications. For this reason, it is better to hold judgment about the battery life difference between Core Duo and Core 2 Duo notebooks until benchmarks are widely available.

 

No, can't agree with abve statement.
1st, I agree Core Duo wins X2 over the edge, not much. The 30 minutes battery life is under 950GM condition vs. X2 Nvidia 6150. In these configuration, you can't say Intel PLATFORM outperform X2 especially on video feature and 3Ds.

2nd, the dailytech article really said Yonah really did a good job on TDP and idle mode that even X2 is hard to catch with. But things get changed in C 2 D, 20W in IDLE mode (not Max TDP) is really a lot. As I said my AMD 3500+ desktop is close to 25+W in idle mode.

3rd, I really don't understand why Intel change their lowest clock into 1 GHz other than original <800MHz in P-M. Maybe there was a barrier or some good reason. Yeah, with higher clock, you will get quicker response, but what do you want it response for when it is in "idle"? 10 mins higher clocking energy vs. <0.3 quicker response? Which one will you buy?

I wish Intel release C 2 D really not like what the article said. But nothing for sure until users fully review it.

 

When I was referring to platform, I was solely referring to it in regard to battery life. Intel, by producing the chipset and wireless card in addition to the processor, has increased battery life in a way that AMD processors + chipset + wireless card cannot match.

I also see where your logic about battery life in Core 2 Duo notebooks come from, but once again, I think we should wait until benchmarks come out because everything before that will be speculation based on paper specs.

And lastly, the original Pentium M had a 800 Mhz speed at its idle state because all Pentium Ms are based on FSB x multiplier to get their Mhz rating. Since the original Pentium M had a 400 Mhz bus, it could use a multiplier of 2 to get that rating while the 1 Ghz speed occurs because the newer Pentium Ms use a 533 or 667 Mhz bus.

 

Anandtech had an article comparing Yonah and Merom, including benches.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2808&p=1

For the lazy, better performance, same power usage. The theory being if you get the job done faster, you give the processor more idle time at the lower voltage states.

 

Agree on the term of PLATFORM that people almost always get confused. When we talk about CD or C2D, we have to clearify if we talk about CPU only or we talke about the platform as whole. Or sometimes better use Centrino certified which implies lot of battery saving. Intel's optimization works great at this level, and that's why I like P-M more than C 2 D.

This could be understand as the defact of FSB when it compare to HyperTransport. Well, finally AMD manage their CPU at lower clock in idle mode, but Intel can't handle it. As long as Intel put great effort to reduce 1GHz power consumption, their battery life will always short. Besides, the new 965GM isn't doing good on its power also at 20+W (paper report). I really don't like to see Intel lose its control of power saving, but it really happen.

True if your idle state is clocked at 766MHz. But 1GHz is close to 25% energy wasting constantly.

 

If you are doing very intensive stuff on battery, then your statement might be true. But often in real world use, the average user when on battery is most likely either surfing the web, using word, listening to music or watching a movie, these tasks do not require the full power of virtually any cpu currently available in the market and I dont think you will ever see the sort of benefits. Most likely the processor will spend most of its time at 800MHz/1GHz and the if you look around reviews for the Turion, it does better in such circumstances than even a Pentium M.

The Turion X2 actually has lower power consumption than the single core Turion. The CD on the other hand has increased power consumption over the Pentium M. This is easily reflected in their TDP. The Pentium M 533MHz FSB was 27W whereas most Core Duos are 31W. This with the introduction of the 65nm process. The Turion ML was rated at 35W, the X2's (still manufactured on a 90nm process) are rated from 31 - 35W. So the X2 is cooler running and gets the same battery life as the single core Turion.

Just a small correction on the bus speed and multipliers, since the bus is quad pumped, the multipliers are dependant on the real bus speed which for a 400MHz QP is 100MHz. So the multiplier for 800MHz will be x8.