Log in

View Full Version : Intel to Use Model Numbers, Not Mhz


Jonathon Watkins
03-21-2004, 07:00 AM
<div class='os_post_top_link'><a href='http://news.com.com/2100-1006_3-5174895.html?tag=nefd_top' target='_blank'>http://news.com.com/2100-1006_3-517...ml?tag=nefd_top</a><br /><br /></div>'Megahertz myth' is a phrase that various folks (AMD & Apple especially) have been bandying around for a while. About two years ago AMD changed from rating their CPUs by actual Mhz, to an equivalent rating, and then recently moved to model numbers for the Opterons (AMD-64 processors). Well, in May, Intel is also planning to change the marketing for most of their line-up to a chip-numbering system. <br /><br />"Intel later this year will adopt a new system for differentiating its processors that de-emphasizes the widely used gigahertz, or clock speed. Sources familiar with Intel's plans said that the chipmaker in May will begin affixing each of its new processors with a number designed to help consumers decipher how the features stack up against other processors in the same family. Intel will use numbers in the ranges of 300, 500 and 700, similar to the model numbers BMW uses on its sedans." So, you'll get the high model numbers roaring up close behind you, flashing you with their headlights and beeping on their horns then? Ah – so not exactly like BMWs then. :wink: <br /><br />"The new system is a dramatic change in Intel's marketing approach because it takes emphasis away from using clock speed as a main measure of performance. Instead, the system will strive to create a scenario in which a person choosing between several 300 series chips, for example, equates the decision to an exercise in choosing a good, better or best processor, sources familiar with the plan said." So then, one best processor to go, hold the Celeron. :lol: <br /><br />"Intel's Dothan Pentium M chips will be grouped inside the 700 series, the sources said. Its 500 series will include both desktop and mobile Pentium 4 chips, and its 300 series will include desktop and mobile Celeron chips. The company is most likely to label each chip with its family name, such as Pentium M, first. Its processor number, such as 700, will follow. Then Intel will list clock speed, cache size, bus speed and other features that impact performance. A chip with even only one different feature, such as a slower or faster bus speed than others, would likely get a different number in order to set it apart".<br /><br />With a mobile Pentium M at 1.4Ghz beating a Pentium IV at many tasks at 2Ghz, it's not surprising that people can get confused. AMD really do seem to be making the running at the moment in terms of initiatives. Intel have adopted 64 bit x86 processing and now this model numbering methodology. Does the model system number make sense to you guys, or do you reckon we will be soon be pining for the days when relative CPU performance was 'simple' to understand? :?:

Kati Compton
03-21-2004, 07:18 AM
Ah, but they say they're still going to list clock speed, just after the model number.

Maybe at some point the model numbers will be comfortable, but I'd think for a while it'll just be one MORE number to remember, adding complexity, not removing it.

ctmagnus
03-21-2004, 07:20 AM
"That thing got a Hemy?"

Aerestis
03-21-2004, 07:26 AM
I don't get it!

Zack Mahdavi
03-21-2004, 08:49 AM
Hmm, I think Intel pulled itself into this dilemma. Their constant struggle to get faster "MHz" only made the fact they can't make the clock speed faster their own problem.

Instead of trying to play around with pipeline designs, etc, to make the processor operate at a higher clock speed, Intel should have worked on building faster processors without focusing on "higher clock speeds."

It's clear clock speed doesn't matter anymore... too bad Intel didn't see the vision until recently... :bangin:

Should Intel adopt a new processor naming system? No... well, not in the way they want to. If every manufacturer adopts a different way of ranking their processors, consumers will be confused. How will they be able to compare performance across brands? They'll end up comparing the clock speeds. The processor giants (Intel, AMD, IBM, Motorola, Sun, etc) need to get together and map out an industry-standard naming convention. Only then will consumers move away from what they're comfortable with, clock speeds.

carphead
03-21-2004, 09:29 AM
Hey man I'm a strictly Intel 700 series person. I wouldn
t dare be seen driving a 500 and don't even get me started on the 300!

Will they come in a range of colours and trim options? I want mine to have a leather interior.

Can I get furry dice with that?

Blue Zero
03-21-2004, 09:41 AM
...So there will be chips like, Pentium SX34.2R EX+
Thats too confusing

Heck, Confusion is what makes life fun :multi:

enemy2k2
03-21-2004, 10:31 AM
These days the upper end of P4s dissipate over 100 watts, and this will only get worse as they attempt raise the clock even further. Even with the new Prescott's 90nm process things have not gotten better. All this means much larger heatsinks are needed and most importantly faster spinning more powerful fans. So not only is a lot of heat produced, but a dang lot of noise. It's the latter part that bothers me most and probably a lot of others out there as well. Next box I build I'm going to give serious thought to water cooling if the situation doesn't get much better.

daS
03-21-2004, 05:25 PM
I like the idea, although I think it would be better if the PC industry press made benchmark data easier to find.

For example, I was in the market for a new notebook a few weeks ago. Speed was important to me because I am using it for heavy database work. Although I knew that the Pentium M gets more out of each clock cycle than a Pentium 4, I never found a good benchmark rating to compair processor models.

guinness
03-21-2004, 05:36 PM
Stupid Intel, now I have to change how I look at AMD's chips now, since they based their ratings on Intel's clock speeds. For example, I'm looking at either upgrading to an Athlon XP 3000+ or a Athlon 64 3000+, now Intel decides to throw a wrench into everything.

retrogamer
03-21-2004, 09:50 PM
MHZ MYTH???? Yeah, of course. But Mhz Myth has been the sole numbering game intel has been playing.. in order to beat their competition. AMD, IBM, and other processor companies has processor running at lower Mhz, but with equal and some times better performance.

OF COURSE MHZ is not a true measure of the performance.. However, sometime like that coming from intel tells you they have hit an wall in pushing Mhz higher and higher..

Recent news of intel latest prescott processor had to be released to run at 2.6 GHz!! that's lower Mhz than previous P4 processors! WHy? it's because power.. the chip would be too hot if ran at higher speeds..

Power issue is a big wall that the entire industry is facing.. Given the amount of resource Intel has, if they can't find a solution.. it's likely the whole industry will be affected heavily by this issue. and forget about moore's law.. this issue essentially can stop everything.

sponge
03-21-2004, 10:27 PM
People made the same end of the world predictions when AMD went to their system, and look at things now: not much of a big difference. All of this will blow over pretty quickly.

ctmagnus
03-21-2004, 11:43 PM
I don't get it!

Sorry for the late reply (had to restore my HD between then and now) but there's the truck commercial (I forget which company) in which the drive-thru operator asks the customer if the customer's truck has a Hemy engine.

pacemkr
03-22-2004, 06:01 AM
The heat is probably the limiting factor in modern processors. However, I think that solutions to this are just around the corner. Necessity is the mother of invention...

A friend recently sent me a link discribing a new, relatively (to watercooling) low cost, cooling system.

Here is the link. Looks very promissing.
http://www.ac4g.com/products/ac4g001.html

Going back on topic. I dont care much of the new naming system. I dont care if the processors are called AIJ@#&lt;DAJJI>112.34.||||.12, simply because we are past entering the stage of confusion. The confusion is here. I've built several systems and every time I had to refer back to the real life benchmarks available on the net. There is no way to know which processor is faster without these. The model numbers mean squat already. Average consumers probably dont know that if they get P4 2.4C they can easily set it to 2.6-3.0 ghz and it is going to be a much cheaper way to get the same speed. And please dont call this overclocking. I am willing to argue that 2.4C and 2.6C are exactly the same processors. The 2.6C is set to higher clock speeds by default thats all. Thus average consumers are bound to pay more. And this model number bussiness is just another way to achieve this.

Racky
03-22-2004, 11:27 AM
Way back in '97 I build an embedded computer using a prototype asynchronous processor developed by Manchester University...

At the time I new they'd never stand a chance marketing it to a MHz obsessed public/industry, despite the power saving advantages. Hell, how do you market a processor which doesn't even have a clock??? :frusty:

Maybe things are changing... It's about time! hehe (subtle pun intended)

There's a picture and schematic on my website if anyone's interested - have a look under projects.

Jonathon Watkins
03-22-2004, 12:48 PM
Cute little robot you built there Simon.

Welcome & good to see you. 8)

Yup, the times are a' changing indeed.

I'm just not sure about remembering model numbers in addition to all the other info out there. My brain can only remember so many numbers at any one time. :lol: