128GHz CPU by the year 2011

[videoclone]

Its crazy what these people thought computers would be like in 2011, The link below is very interesting indeed. I like the mid 2011 prediction of a 128 GHz CPU and this line from one of the comments below it

"If 10 GHz is the best that Intel can do by 2011, AMD or somebody else is going to eat their lunch."

http://www.geek.com/articles/chips/intel-predicts-10ghz-chips-by-2011-20000726/

 

[Lonbjerg]

And then physics hit...:biggrin:

 

[gevorg]

They probably confused GHz with GBs.

 

[Soccerman06]

LOL those posts make me sad

 

[Mr. Pedantic]

Hooray for an ignorant public...?

 

[AznAnarchy99]

LoL

"Get a loan from your bank and get an Apple. With Apple you turn it on, start your application and start becoming productive. Once you become more productive you'll make more money. Take that money and repay the loan. After your loan is paid off, trade in your Ford Fairmont for a new Explorer. "

 

[videoclone]

Ford Explorer lol

 

[Ancalagon44]

Oh geez check this Apple fanboy:

doesn't matter the speed of intel's chip in 2011 because motorola and ibm will already have chips out by then that are more effective, use energy better, and run applications ready for the new iMac's that are introduced at the 2011 MacWorld. They'll run at somewhere near 7GHz and still be faster than an 11GHz or even 128GHz sh*t that intel puts out.. - by StickWithApple

I would love to get a hold of him now and rub this quote in his face.

 

[nanaki333]

i guess if you add all the ghz in a quad 16 core BD when they finally launch, it's SOMEWHAT true

 

[Idontcare]

They probably confused GHz with GBs.

LOL, close, they entirely misunderstood what Moore's Law says/is about.

If not, Moore's Law states that the number of transistors in a common microprocessor will double every 18 months. As well, it can be applied to processor speed and many other computing/technology metrics.

Uh, no, it was never intended to be applied to clockspeed. It was also never intended to be applied to any performance metric of any kind.

Moore's law is quite explicit in that it defines the minimum point on the manufacturing cost per component as reaching lower minimums when the component per chip were roughly doubling every year.

Moore's law is about manufacturing cost and the number of components on an IC.

That's it, nothing more, and yet that alone has significant ramifications to the future of Moore's law because Moore's law has always been about the scaling of production/manufacturing costs and it is those costs which are scaling to unaffordable heights because of mask set costs and litho requirements.

 

[soccerballtux]

Wow!!! Yet another ground breaking Moore's law post!!!

here's a Godwin's Law equivalent. You all can thank me later!!!

err, meant to throw in a reference to CPU Frequency in the "length of discussion" part...

 

[AdamK47]

I remember reading this AnandTech article back in 2000 before the Pentium 4 was released. Intel had high hopes for Netburst.

http://www.anandtech.com/show/680/6

 

[mrjoltcola]

Off-topic but the most hilarious part of that thread was the teenage punk bragging about how his uncle was VP of Ohio Dept of Transportation, and that is how he got his AutoCAD "legit like", then he gets _spanked_ by the guy from AutoDESK's Anti-Piracy group 3 posts later.

Pure, classic entertainment!! Wiping coffee off my screen.

Game Junkie? Preteen? Excuse my Mr. Bill Gates wannabe, but where I come from thats called insulting. If you cant think of anything worthwhile to say other than talking smack, then dont. And by the way, I didn't get 3D Studio Max and AutoCad 2000 from some gay warez site. My uncle happens to be vice pres. of ODOT. (for you intelectually challanged, that the Ohio Department Of Transportation. I can get anything I want, basically

Yeah the big boys are the people that can actually pay for their software instead of pirating it from family members. It's pukes like you that make our stock price drop because you can't pay for software you want to use. Looks like I will have to look into the number of AutoCAD seats ODOT has paid for.

Signed
Jess Owens
Anti-piracy Group
Autodesk, Inc.

PS and if you think that I am kidding see if you get an email bounce from
jess.owens@autodesk.com

I can hear the phone call to ODOT...

Yes sir, good morning, this is AutoDESK, how are you this morning?

*Fine thanks*

Yes sir, we were just auditing your licenses...oh by the way, how's that whiz-kid nephew of yours?

...

 

[Mopetar]

LOL, close, they entirely misunderstood what Moore's Law says/is about.

Uh, no, it was never intended to be applied to clockspeed. It was also never intended to be applied to any performance metric of any kind.

Whether or not it was intended to doesn't really matter as, interestingly enough, it does serve as a "good enough" predictor of performance. Also there was a recent Slashdot article that discussed another spin on Moore's Law whereby the amount of power required to compute something halves every 18 months.

As much as it gets abused, it's not bad for making rough guesses about future performance or energy costs. Also, it's not like we haven't had anyone producing 100 GHz capable hardware. It's just that it's not feasible to produce them at this point and time.

 

[Mopetar]

Off-topic but the most hilarious part of that thread was the teenage punk bragging about how his uncle was VP of Ohio Dept of Transportation, and that is how he got his AutoCAD "legit like", then he gets _spanked_ by the guy from AutoDESK's Anti-Piracy group 3 posts later.

Pure, classic entertainment!! Wiping coffee off my screen.

I can hear the phone call to ODOT...

Yes sir, good morning, this is AutoDESK, how are you this morning?

*Fine thanks*

Yes sir, we were just auditing your licenses...oh by the way, how's that whiz-kid nephew of yours?
...

That's pretty damned funny. Reminds me of this if you want another good laugh.

 

[mrjoltcola]

That's pretty damned funny. Reminds me of this if you want another good laugh.

Like it like, good stuff.

 

[Idontcare]

Whether or not it was intended to doesn't really matter as, interestingly enough, it does serve as a "good enough" predictor of performance. Also there was a recent Slashdot article that discussed another spin on Moore's Law whereby the amount of power required to compute something halves every 18 months.

As much as it gets abused, it's not bad for making rough guesses about future performance or energy costs. Also, it's not like we haven't had anyone producing 100 GHz capable hardware. It's just that it's not feasible to produce them at this point and time.

My point is that just because you can fit some characteristic relevant to technology to a linear fit on a log-normal plot does not make that thing suddenly become an extension of Moore's Law.

Moore's Law is specific. Yes other aspects of the technology in this industry can also be modeled with log-normal linear fits, but that doesn't make it "Moore's Law sayz...".

My toyota sienna mini-van is an automobile, not all automobiles are toyota sienna mini-vans.

 

[DeathReborn]

I guess we're still waiting for IBM to figure out how to put those 350-500GHz transistors they have into a complex chip, 5 years & nothing new to report.

 

[Idontcare]

I guess we're still waiting for IBM to figure out how to put those 350-500GHz transistors they have into a complex chip, 5 years & nothing new to report.

^ hell, I think at this point their customers would be happy if they just managed to squeeze out the chips with HKMG they were talking about 4-5 yrs ago.

 

[Lotheron]

Like it like, good stuff.

That *smells* fake.

Just saying.

 

[tweakboy]

Wow, good find. I remember something like that back in 1998. When we were @ 450Mhz lol P3 600Mhz. They guessed they would go to certain Mhz, but turns out Mhz is not what makes the speed of CPU, its number of things.

I dont see 10Ghz for a long long time. 2018 maybe..

 

[mrjoltcola]

That *smells* fake.

Just saying.

You mean the bash.org post? Yeh.

A lot of stuff on bash is, but still funny. The Autodesk guy from the OP's original thread sounds authentic, though.

 

[Mopetar]

My point is that just because you can fit some characteristic relevant to technology to a linear fit on a log-normal plot does not make that thing suddenly become an extension of Moore's Law.

Yup, and mine is that there are a lot of other useful things correlated with Moore's law. Anyhow, back on topic before this just gets reduced to silly pedantry all around

The funniest part about having chips that can clock 300 - 500 GHz is that you start running into constraints about how far apart certain features can be located as the electrical signal can only travel so far in a given cycle. Not necessarily a problem if you're also making really small chips, but still something to keep in mind.

 

[ThatsABigOne]

Wow, good find. I remember something like that back in 1998. When we were @ 450Mhz lol P3 600Mhz. They guessed they would go to certain Mhz, but turns out Mhz is not what makes the speed of CPU, its number of things.

I dont see 10Ghz for a long long time. 2018 maybe..

Mark my words. That will never happen. It is all going towards parallelization.

 

[Mopetar]

Mark my words. That will never happen. It is all going towards parallelization.

That still doesn't mean we won't get a 10 GHz CPU in the future. Eventually you hit as much of a wall with parallelization as you do when trying to ramp up the clock speed. It might be a while before we get there, but sooner or later, increasing clock speeds is going to be the path of least resistance to get better performance.