P4 will scale above 10GHz

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.

Raystonn

Distinguished
Apr 12, 2001
2,273
0
19,780
http://www.intc.com/pressroom/archive/releases/cn121100.htm

"Intel Corporation researchers have achieved a significant breakthrough by building the world's smallest and fastest CMOS transistor. This breakthrough will allow Intel within the next five to 10 years to build microprocessors containing more than 400 million transistors, running at 10 gigahertz (10 billion cycles per second) and operating at less than one volt."

-Raystonn

= The views stated herein are my personal views, and not necessarily the views of my employer. =
 

TheAntipop

Distinguished
Dec 31, 2007
1,315
0
19,280
would you really want a 10ghz processor based on 7th generattion processing? hell we might be seeing the first consumer quantum computers by the time a 10 ghz processor comes out. and then theres clockless computing. 7th generation wont last long enough to see 10ghz, nor would anyone want it too. because if youve noticed, speed increases are beginning to have diminishing returns.

<A HREF="http://synthetic.bsd-fan.com/pika.swf" target="_new">Hyakugojyuuichi!!</A>
 

Raystonn

Distinguished
Apr 12, 2001
2,273
0
19,780
"speed increases are beginning to have diminishing returns"

On the old P3 and Athlon cores yes. This is why a paradigm shift to a whole new core was in order. 10GHz processors based on this same P4 core will be available in about 3.0 - 3.5 years. This follows Moore's law.

-Raystonn

= The views stated herein are my personal views, and not necessarily the views of my employer. =
 

TheAntipop

Distinguished
Dec 31, 2007
1,315
0
19,280
first of all did you read my entire post? do you know what a 7th generation processor is?
also i took every single benchmark (except memory performance) from toms last review of the the p4 1.7ghz and compared them to the 1.5 ghz values in the same tests. While the 1.7 is 13.3% faster on paper, the average speed increases on the benchmarks was only 7.9%. thats barely more than half the return in speed increase. the althlon 1.2ghz vs. 1.3ghz fared much better. with only an 8.3% increase the average speed increase was 7.1%, much much closer to its paper speed. How do you explain these results away? If youd like I will repeat the process for 1.4->1.5 and 1.3->1.5, maybe that will make you realize what "diminishing returns" really are.

<A HREF="http://synthetic.bsd-fan.com/pika.swf" target="_new">Hyakugojyuuichi!!</A>
 

Raystonn

Distinguished
Apr 12, 2001
2,273
0
19,780
That has nothing to do with diminishing returns. You are comparing two different processors that have a different number of average instructions per clock. It's well known that the P4 does not do as much in one clock as a P3 or Athlon does. However, total performance is measured in clockspeed multiplied by 'average instructions per clock', not clockspeed alone. Equating clockspeed to performance is a complete misnomer. I might as well design a 1Hz CPU that has one clock per second. I can fit an awful lot of instructions in that one clock, can't I? I could completely overwhelm any current processor on a clock per clock comparison.

As far as benchmarks go, I'd wait until software that has been optimized for the P4 comes out before trusting them. Currently they hold no credibility.

-Raystonn

= The views stated herein are my personal views, and not necessarily the views of my employer. =
 

TheAntipop

Distinguished
Dec 31, 2007
1,315
0
19,280
i was comparing two p4 processors to each other. are you saying that the p4 1.7ghz and the p4 1.5ghz perform different number of instructions per clock? wouldnt that make the p4 1.7ghz processor much more than 13.3% faster than the 1.5? it seems you are trying to play word games here to cover your ass, but it aint workin. the fact remains the same: higher clocked 7th gen processors are starting to show diminishing returns. next gen processors are coming out next year, and another generation will more than likely pass before the p4 hits 10ghz. so would you want a 10ghz 7gen processor or a 10ghz 9gen processor? its like asking if you want a 1ghz pentium or a 1ghz pentium3, the difference is obvious.

<A HREF="http://synthetic.bsd-fan.com/pika.swf" target="_new">Hyakugojyuuichi!!</A>
 

Raystonn

Distinguished
Apr 12, 2001
2,273
0
19,780
Actually the difference is more like comparing the Pentium Pro with the Pentium 3. They use the same basic core. That's what will be happening with the new Pentium 4 core. It will scale all the way to 10GHz without much difficulty. There will probably be new 'generations' of processors using said core though.

-Raystonn

= The views stated herein are my personal views, and not necessarily the views of my employer. =
 

Ncogneto

Distinguished
Dec 31, 2007
2,355
53
19,870
<font color=purple> "Geeze calm down you AMD gimps need to know when youve been beaten and beaten bad. "</font color=purple>

What was the memory bandwith on the p4 again? LOL not only is the p4 getting its but kicked in the desktop arena lookout intel they are going to have AMD laptops that whoopya too!
<font color=red>
Sources with knowledge of the new chip claim it will have the industry's highest mobile processor clock rate and a 4.2Gbit/s memory speed, exceeding even the fastest desktop PC on the market today.

Details of Palomino leaked to EBN describe a chip with a clock rate of 1.3 to 1.4GHz, supported by double-channel PC2100 memory modules with 4.2Gbit/s data rates. </font color=red>

That would be the palomino Spud, this quarter! Dual channel DDR in a laptop.

Do you know the muffing man, the muffin man,.............



A little bit of knowledge is a dangerous thing!<P ID="edit"><FONT SIZE=-1><EM>Edited by Ncogneto on 04/28/01 04:42 AM.</EM></FONT></P>
 

girish

Distinguished
Dec 31, 2007
2,885
0
20,780
silicon has its limits, and P4 architechture however "scalable" cannot reach 10 GHz, at least in the silicon era as the only one we know of now.

remember a few years ago physical limitations of copper on a PCB were realised and the system bus stuck at 66~133 MHz and CPUs started running faster internally on some multiple of the system bus. similarly, limitations of silicon are being reached, it would be replaced by some other material.

not that we wont see and mutli-GHz machines in recent future, i dont think they would be P4. newer material will need a new architechture that might be named as P5 or higher.

now, is 400 MHz bus of P4 really 400 MHz, its Quad-pumped Dual channel RDRAM 100 MHz bus that is *scaled* to 400 MHz by some innovative techniques. similarly, future processors may have more than one processor units inside that will perform parallely, a sort of multiprocessing withn a single chip - and that will scale its performance, say quad-unit P4 each running at 2 GHz having 400 MHz memory bandwidth could be labelled as 8 GHz processor having 1600 MHz memory bandwidth!!!

the VLIW (Very Large Instruction Word) processing is not far away from this...

<font color=blue>die-hard fans don't have heat-sinks!</font color=blue>
 
G

Guest

Guest
I do know the muffin man why???

SPUDMUFFIN

<font color=red>Being Evil Is Good. Cause I Can Be A Prick And Get Away With It.</font color=red> :lol:
 

Phelk

Distinguished
Apr 21, 2001
203
0
18,680
Girish...well done. I knew it would take someone from outside of the US to raise some real points. 10Ghz will not be reached by squeezing more juice out of the P4, a single chip (3 * 3.33Ghz) P4 will effectively give 10Ghz much cheaper and earlier (And will make the marketing guys happy). AMD will be stacking chips once they get to 0.13micron, thats why they built Dresden the way they did.

The CPU game is changing, its called 64bit for massive scale and parallel processing for power. There aren't many applications looking forward that will run noticably better on a 10Ghz than 2 * 5Ghz.

Kiwi in the UK
 

TheAntipop

Distinguished
Dec 31, 2007
1,315
0
19,280
Girish...well done. I knew it would take someone from outside of the US to raise some real points
actually that point was already raised much earlier.

<A HREF="http://www.512productions.com/lobstermagnet/" target="_new">Hyakugojyuuichi!!</A>
 

kal326

Distinguished
Dec 31, 2007
1,229
108
20,120
Well wasnt the Pentium 60 the best back in 1993. So accordingly then the Pentium 240 should have been out in 96 and the 960 in 99. So in 2002 we should have 3840 chips. That is if the law is exact. (User is bored)

Somebody call Guinness. I'm about to go zero to drunk in <b>twenty dollars!</b>
 

khha4113

Distinguished
Dec 31, 2007
2,143
0
19,780
There is an article for anyone who might be interested.
http://www.ibm.com/news/2001/04/27.phtml

<b>IBM scientists have developed a breakthrough transistor technology that could enable production of a new class of smaller, faster and lower power computer chips than currently possible with silicon.
The researchers built the world's first array of transistors out of carbon nanotubes -- tiny cylinders of carbon atoms that measure as small as 10 atoms across and are 500 times smaller than today's silicon-based transistors. The breakthrough is a new batch process for forming large numbers of nanotube transistors
</b>
 

killall

Distinguished
Dec 31, 2007
979
0
18,980
a foster will be a P4 in the same way that a P1 60Mhz is the same as a P3 1000Mhz

where did all my money go, ahh... [-peep-], forgot that i lived in london
 

girish

Distinguished
Dec 31, 2007
2,885
0
20,780
well, the technology is still in the labs and it will take a while for it to mature, and be mainstream.
meanwhile, processors have to bear with the current silicon technology that hsa already almost at its end.

now processors will NOT increase in MHz as much, but WILL increase in performance that better MHz would deliver, that newer techniques should be developed to get more out of same MHz.

after a while, these carbon nanotubes may takeover, or who knows optical transistors may take over and processors may see TeraHertz speed.

But is P4 architecture suitable for it...?

too bad you purchased a silicon based mobo right away and you cant upgrade your processor to carbon :-(
... thats at least 10 years away pal!

girish

<font color=blue>die-hard fans don't have heat-sinks!</font color=blue>
 

SammyBoy

Distinguished
Mar 16, 2001
689
0
18,980
Could I just add that Moore's Law states nothing about speed. What it states is "Every 18 months, the number of <b> transistors </b> on a processor will double." The number of transistors is not the only thing that affects the MHz/GHz of a processor. Look at the GeForce3. It has 50 million or so transistors. The Kyro2 has 15 million, and it is not two-thirds slower than the GeForce3. In fact, in some cases, it runs at almost 90% of the GeForce3. The nanotube technology is what IBM and others have been trying to get figured out so that they can squeeze more and more transistors onto a chip. Now that it can be done in lab, and results in good yields, it will be at least five years before it is out in the fab plants, just like EUV. The foster core may hit 10GHz someday, but it will be a cold day in hell if it is the P4 that does it. And hell, I say silicon chips in general are dead. Bring on the bioware!

-SammyBoy

Without Evil, there can be no Good. Therefore, without an Intel, there can be no AMD.
 

TheAntipop

Distinguished
Dec 31, 2007
1,315
0
19,280
very good point, i didnt know that and im sure a lot of other people didnt as well. thanks man.

<A HREF="http://www.512productions.com/lobstermagnet/" target="_new">Hyakugojyuuichi!!</A>
 

girish

Distinguished
Dec 31, 2007
2,885
0
20,780
well, i agree to almost all of it...
but every advancement in technology has its implications as well. doubling the number of transistors not only means smaller chips, but lower power consumption and other physical effects like inductance and capacitance as well, and thereby ability to produce faster chips.. how fast, its just not a measure of MHz, but also the organisational improvements that smaller circuits allow to pack into the same area.

inclusion of cache, and the L2 cache on the chip itself allowed to increase the performance, the barrel shifter introduced in the 386 allowed multi-bit shift operations to be done in single cycle, increasing the number of pipes in the superscaler pentium which were actually two independent execution units almost doubled the performance at the same MHz as that of the 486.

this, i guess, could be the real meaning of Moore's law. and it should not apply to just silicon, newer processors made with newer materials, newer techniques and newer architechtures might as well continue to obey this legendary law.

girish

<font color=blue>die-hard fans don't have heat-sinks!</font color=blue>
 

Raystonn

Distinguished
Apr 12, 2001
2,273
0
19,780
http://www.intc.com/pressroom/archive/releases/cn121100.htm

"Intel Corporation researchers have achieved a significant breakthrough by building the world's smallest and fastest CMOS transistor. This breakthrough will allow Intel within the next five to 10 years to build microprocessors containing more than 400 million transistors, running at 10 gigahertz (10 billion cycles per second) and operating at less than one volt."

With this technology we will see the P4 core scale upwards of 10GHz. After that, it's anybody's guess where we go.

-Raystonn

= The views stated herein are my personal views, and not necessarily the views of my employer. =
 
G

Guest

Guest
That is incredible truly incredible. Tell you one thing Intel is on the ball this time. Two thumbs up from me maybe the monkey's will too.

SPUDMUFFIN

<font color=red>Being Evil Is Good. Cause I Can Be A Prick And Get Away With It.</font color=red> :lol:
 

TheAntipop

Distinguished
Dec 31, 2007
1,315
0
19,280
where in that article does it say the P4 will go to 10ghz? sure they have the technology to go that high, but i cant see it happening without a heavily modified P4 if you could still call it that. Do you know how hot a P4 core would be running at 10ghz? Do you even consider this kind of stuff before posting? Do you know how a processor is built and how it functions? Jesus man, with every post you sound even more stupid than before. Quit trying to sound technical because you obviously arent.

<A HREF="http://www.512productions.com/lobstermagnet/" target="_new">Hyakugojyuuichi!!</A>
 
G

Guest

Guest
IBM is also working on photon switchs for cpu's. Almost everyone is working on it but IBM seems to be closest to the goal, products 3 to 5 years from now.

Anim88tor
 

Raystonn

Distinguished
Apr 12, 2001
2,273
0
19,780
"where in that article does it say the P4 will go to 10ghz? sure they have the technology to go that high, but i cant see it happening without a heavily modified P4 if you could still call it that. Do you know how hot a P4 core would be running at 10ghz? Do you even consider this kind of stuff before posting? Do you know how a processor is built and how it functions? Jesus man, with every post you sound even more stupid than before. Quit trying to sound technical because you obviously arent."

Wow, you're easily angered. Take a deep breath. Might I suggest an anger management course?

If you'd take a moment to look at the URL I pasted in the very first message of this thread, you'd see the reference where it is stated that the P4 core will go beyond 10GHz. I never make up anything I write. Before you start making personal attacks against those you know nothing about, it would be best to check your sources. (Not that having sources that contradict someone's argument is any reason to attack them personally. You should always attack the argument, not the person. I'm just pointing out that you have no sources anyway, and lack any kind of credibility here.)

-Raystonn

= The views stated herein are my personal views, and not necessarily the views of my employer. =
 

Ncogneto

Distinguished
Dec 31, 2007
2,355
53
19,870
"Intel Corporation researchers have achieved a significant breakthrough by building the world's smallest and fastest CMOS transistor. This breakthrough will allow Intel within the next five to 10 years to build microprocessors containing more than 400 million transistors, running at 10 gigahertz (10 billion cycles per second) and operating at less than one volt."

With this technology we will see the P4 core scale upwards of 10GHz. After that, it's anybody's guess where we go."

Good article. however nowhere does it mention that that technology would be implemented in the core of the p4, only that processors up to ten gig would be obtainable in 5 to ten years. Now, by using moore's law that you like to use lets see. Assuming that we will be at 2 gig in a couple of months.......

in two months........2 gig
18 months....4 gig
36 months....8 gig
54 months....16 gig
54 months/12 = 4.5 years

Point being Moore's law is going to no longer be effective soon.

Now what is your perspective on the belief that others in this thread seem to think that there motherboard may actually be usable all the way up to ten gig? And what is intels plan for upgrading the pci ide bus's? What have they got to bring to the table in the lines of hypertransport tech ( AMD) Which I may add is soon to be used in the x-box and will thus most likely be very well supported by M$.

A little bit of knowledge is a dangerous thing!