Dual Core Notebooks - Page 3

i thought intels tdp was calculated at 70% use?

First off let me go on the record by saying I do not work for Intel, and if I did I certainly would not be shamed of it. The avatar and signature were just that, nothing more.

I mentioned that fact about it was in development for a few years, because I took your statement of a stop gape as Intel rushed Development and slapped to Dothan’s together. I was making the point that it is not a stop gap product.

I was merely pointing out facts towards your statement of Yonah being not so hot in the performance department. What I pointed out was improvements over Dothan, to show that it’s not close to two Dothan’s slapped together. I do agree with you on everything you said. By no means did I want to make it sound like Yonah is going to be this amazing beast with the power of two Pentium 4’s @ 7GHz. I agree Merom will easily beat Yonah in every area but power usage.

About the power draw of Yonah and Napa it self. Napa as a whole platform will draw and use less power overall the Sonoma. Yonah as its rated speeds for laptops will have a TDP of 31W. The reason for the power figures is to basically cover up the higher power draws of Yonah, and most importantly Merom. Saying that the 50W rating I am pretty sure is for a desktop version of Yonah that will be clocked higher, as Intel does not need to worry about power restraints. I don’t know much about the plans Intel has for Yonah on the desktop, but they have mentioned their server plans for it in the form of Sossaman. The reason I said about Merom is because as of now Merom’s TDP is 45W. Certainly not light on the power foot print, but it will hold over until Intel changes over to its 45nm fabrication process.

WTF? i thought merom was going to have super low power use and 8 hour battery life?

The last real explanation that Intel used was from a whitepaper for the release of the Williamette P4. It stated that TDP was set at a level slightly higher than the maximum realworld sustained power usage found with a suite of CPU intensive applications, which corresponded to about 75% of the maximum theoretical or instaneous power usage of the Williamette P4. This number changed to around 85% with Northwood.

As for the power consumption of Dothan:

http://babelfish.altavista.com/babel...tops/page4.php
Here, the difference between idle, 800MHz/0.988v and running cpuburn 2.13GHz/1.34v is 17W.

You need to reduce the power usage of the other components, like the LCD screen and video card to get that kind of battery life.

here it is,

"
Correction: This story incorrectly stated maximum power consumption levels for Intel's Merom chip for notebooks. Intel has not revealed this information for the chip, due in late 2006. A few years later, Intel expects to release a yet-to-be-named chip with maximum power consumption of 5 watts and an ultra-low voltage version that consumes 0.5 watt."

bastard cnet

.5 watt... we talking about a 286, measured in kilohertz?

Sorry about mistaking you for an IDF rep, it's just that a number of people here thought that you were.

Perhaps we really agree on these things more often than not. What I meant was that Yonah, as compared to Merom from Intel, and socket F K8L (Oct 06) socket F K10 (Q4 07) from AMD for that matter, does not represent a substantial redesign of the CPU core components and will not bring dramatically better performance than its predecessors, with the exception to dual core, which will definitely aid in multi-threaded apps, but as to single threaded applications, I have some doubts as to if Yonah is really going to perform better than Dothan, still have some of the performance constraints inherited from the original P6 core design. That's all I meant, recognizing the number of innovations that Yonah will bring to the table.

This thread has the most crap written in it that I've read in a long time

Ok Vista is not XP with gloss, its a totally rewritten OS hence all the delays and it's based on winserver 2003 and not XP and 2003 is optimized for dual core.

OK and as for another OEM laptop maker releasing dual core by end of Dec besides clevo IT'S NOT GONNA HAPPEN. please let us know who the maker is and give links to the presentation.

OK rant over,
my rant was based on FACTS not hearsay - please post where you guy's are getting your info from. links please. PS if anyone wants links to what I've said then just ask or you can google.

z

vista is not a totally rewritten OS. from what i understand, the early builds were based on windows xp and then the "rewrite" was when they moved it to windows 2003.

and at the end of the day, it really doesnt make a difference. the point is that when vista is released, they will have very little to show for 5+ years of development.

*Sigh*

The core kernel for Windows XP and Windows Server 2003 are close enough that for the most part, they can be considered to be in the same family. (They are theoretically in the same generation, like 2000 Professional and 2000 Server, Advanced Server.) Thus, basing it on one means it's really basing it on both.

Also, you'd think that if there are any performance gains - it'd be pushed to the client level first. Server platforms usually go for stability first, performace a close second.

The only "real" thing that Microsoft did for multi-core is their licensing.

As far as I know, 2003 does not have any specific performance tuning for multi-core designs. With the multiple multi-core designs being pushed out by all the major x86/x64 chip makers, optimizations are hard - how do you handle different cache access designs?

Mind you, I could be completely wrong - but none of the service packs or hotfixes I've seen talked about performance gains in multi-core processors.

the alienwares now have amd processors

Thats the clevo d900k. (Assuming your talking about the Auraura m7700 which can have the Athlon X2 and Athlon FX)


Companies generally like to tell people about stuff they're working on, via reps and such for obvious reasons. If anybody else was developing an amd socket 939 mobo in a laptop with those desktop processors supported we would likelly know. But, the reason for that is not because the technology isnt there... its because thats not where the laptop market is. Most people (Not us... people here are _mostly_ computer people, who actually care about what they buy) don't want a 13 pound brick of a notebook. Alot will, but Dell, Compaq, Toshiba, Sony, the laptops the un-informed buy (Not dissing those makers, just .. yeah you get it) generally sell mid range notebooks. More money there. Its just not as profitable to spend XMillion (or whatever magnitude) dollars to R&D a laptop to dissappate 110 W of heat. Clevo obviously has a good R&D team as they produce laptops left and right. MSI may not have that capacity (Or uniwill etc.).

Is there a dual-core laptop that's small? If not then my first statement about dual-core CPU's requiring large monoliths of a "laptop" was right. The larger amount of heat put out by the dual-core CPU's REQUIRE large laptops. So yes, the laptops are larger. Once they come out with small laptops w/ dual-core, then more people will care. Anyone?

there will be small dual core laptops when intels yonah and amds taylor come out. you cant really expect to put a desktop athlon x2 into a 6 pound notebook.

There are TWO dual core processors on the market right now, the pentium D and the Athlon X2. Both of which are performance chips. They also cost alot. So, taking this into account, we realize that they weren't designed to be put into a laptop in the first place (and IMO a pentium D NEVER should be). So, shoving one into a laptop requires extra cooling solutions etc. thus bigger heavier laptop.


Now, lets look at the common thin and cool laptops. Whats in 'em? Mobile processors. Pentium M, Turion etc. These chips were designed to run cool. So you can get thin and light. You cannot expect to put a chip, which was designed to torch the benchmark tests and outperform any other chip on the market, to be as cool as something which was designed to be cool.

Now, nowhere did i say though that dual core has to be screaming performance. Just because the only 2 dual core chips that are out now are performance chips doesn't mean sometime in the future (which, there will be, as mentioned before) we'll have mobile dual core. Simply put, the current dual cores are performance, thus big laptops, in the future, we get mobile dual cores, and we'll have small laptops. There is no relation whatsoever between dual core and big laptops (big heat etc.), but there is a relation between big performance chips and big heat. (Like the FX series in a laptop etc., and any P4)

Sigh, No operating system is "optimized for dual core". None. Zero.

Every modern pre-emptive multitasking kernel (Linux, Unix, Solaris, BSD, NT, Windows 2K/2K3/XP, BeOS, OSX, Nextstep, VMS, etc etc) that can schedule processes/threads for execution can handle multiple CPUs. Nothing needs to be changed for "dual core" simply because it looks like standard symmetric multiprocessing to the kernel.

Jeez, I've been running dual cpu desktops since win2K come out. Dual core simply makes it cheaper to implement.

FWIW, my gaming systems are NEVER dual cpu. First games typically do not thread very well. Second, the overhead of SMP is there. Third, a single very fast clocked cpu is generally better for games than two slower clocked CPUs.

So early next year is when we should see small laptops with dual-core CPU's?

Not likely

The first dual core Pentium-M (Yonah) is spec'd to dissipate about 50w of heat under load. That would preclude "small" laptops. I expect to see first generation dual core laptops in DTR formats. The second generation, Merom should produce much less heat and possibly support thin and light formats.

This really has nothing to do with dual or single core. Each transistor produced at a given process size (.90 micron, .65micron, etc) and running at a voltage (1.3v for current P-Ms) will produce a set amount of heat. If a dual core chip adds 40% more transistors its heat/power envelope will go up by 40% if all other variables remain the same.

The sad fact is that the majority of users simply do not care. For what most consumer and business users do with their computers 1000Mhz P3 performance is adequate; more memory and faster disk are more important. My work laptop adjust its clockspeed based on cpu load, and I log the changes. Less than 10% of the time does my P-M ever go above 800Mhz running general business apps, and most of the 10% is at boot time.

If I had to choose I'd pay for a 7800go over a dual core processor for my personal laptop simply because the 7800 will have greater impact on games.

Low-voltage yonahs that run from ~15-24w will be launched right along with the 25-49w line. Those will be in the <6 lb category. Those won't be released higher than 1.66 Ghz at first, but I don't think it will be a problem.

I'm sure all the tech sites are eager to benchmark it, especially since intel's worked on the pentium m's fp performance.

well, i want a dual core with PCIe that runs cooler and faster. It looks like I am still 6 months away....so for now, I guess I stay with the Z71v