What's clockless CPUs?

[Guest]

What is a clockless CPU, as described in Raystonn's post about 10GHZ P4? Plus, if a P4 can go up to 10Ghz, wouldn't AMD and Cyrix do so to?

Oh this one's addressed to Raystonn himself personally. Where did you learn all your knowledge from? You can quote many stuff and I understand most of them. But as Antipop said, I am not so sure how you get your knowledge. If what u say is bluff, then he's right. Your using techy terms to cover your ass up, but if your right, well maybe I guess your really that good in hardware.

We put text in here and that's our signature?

 

[Guest]

Search the Technews for "clockless" on Tom's site. There are a couple of articles on clockless CPUs.....

"<font color=blue>Intel</font color=blue>, <font color=red>You can't spend more but you can buy better</font color=red>...."

 

[TheAntipop]

i dont really doubt that he knows what hes talkin about, because the information is so widely available and cpus themselves are pretty simple in design (alhthough all the newer additions complicate things, but at the core, its still really easy to understand).

a clockless cpu, as ibm is now pushing, basically unlocks the cpu from the limitations from having to do all its calculations based on a cycle. lets say for example, that a cpu is passed a really complicated flop and a relatively simple integer op. it will take 1 whole cycle to complete the float op, but only part of one to complete the integer, so all those other integer instructions behind it will have to wait while nothing is done in that area. a clockless cpu lets the cpu go true full throttle all the time in all aspects of the processor, wihtout being held down by timing mechanisms. in theory it a lot faster, but in reality it may never make it because of all the marketing implications (i mean honestly how does the average person tell how fast a computer is? do you think they read hardware review sites?). personally i think that by the time clockless cpus could get out into the market we could be seeing quantum computing emerge on the extreme high end of computers, which i look soooooooooo much forward to. mmmmmmmmmmmmmmmmmmmmmmmm, qubits.

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

 

[Raystonn]

"if a P4 can go up to 10Ghz, wouldn't AMD and Cyrix do so to"

Not with their current designs. They would need a whole new core, as Intel has created with the P4.

"Where did you learn all your knowledge from?"

I've been in the industry for a great length of time and have years of research under my belt. I don't usually talk about myself because who I am has nothing to do with my arguments. I always provide references so arguments can stand on their own.

-Raystonn

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

 

[Guest]

>>>>>"but in reality it may never make it because of all the marketing implications (i mean honestly how does the average person tell how fast a computer is? do you think they read hardware review sites?)."<<<<

Ahhhh, joy, it will be in operations per second. You know how the the Geforce does it, Nvidia says they had the first Giga (million) texel processor, well the cpu's of the distant future will be measured in millions of operations per second! It'll be awesome...Total realism in games (which could be good or bad, however you look at it.)

Aklein

Life is hard...Live with it.

 

[Raystonn]

"Giga (million)"

Giga would actually be billion.

"cpu's of the distant future will be measured in millions of operations per second"

This is essentially the same as 'millions of instructions per second', or mips. CPUs have been measured by such in the past. The problem is every instruction takes a different amount of time to complete. There are also branch mispredictions and such that can decrease the average instructions per second. Making an accurate measurement becomes very difficult.

-Raystonn

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

 

[Grizely1]

Yes, Giga is billion, mega is million. Hehe soon we will be measuring in GIPS! ROFL

---------
I am the first and only one with a 16MB GeForce2 GTS graphics card!

 

[TheAntipop]

how well do you think people will adjust when all of a sudden they see a 1 gips processor? theyll think "hell i gotta 2 ghz processor, what do i need this POS for?" im just sayin it will be a hard sell considering it will be the first major change in computer hardware in a LONG LONG time.

this bring me to quantum computing yet again...

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

 

[HolyGrenade]

I saw a programme on .tv the other day. It showed a 0.03 micron core they're working on. The intel white coat guy said it will be in production at around 2010.

You were saying the p4 core will reach about 10GHz around 2006? (I cant be bothered to go to the other thread to find out), How small do you think the core will be around that time?

Also, saying that cyrix and AMD cannot reach that clock frequency with their current core may be true but it is misleading in a way. Especially because the Athlon has always had a far better performance to clock ratio than the P3 and definitly the p4 core.

The P4 by intels own admission was designed to "dramatically" increase the clock speed of the core. The actual performance increase is another thing.

This post may be directed at you, but the information is mainly for other people. I think you know all of this already, but there are lots of people not familiar with the actual relation between clockspeed and performance, and the role it plays in the cpus of different companies.


<font color=red>"My name is Ozymandias, King of Kings:
Look on my works, ye Mighty, and dispair!"</font color=red><P ID="edit"><FONT SIZE=-1><EM>Edited by holygrenade on 05/03/01 09:05 AM.</EM></FONT></P>

 

[Raystonn]

As performance is the product of clockspeed and IPC, CPU performance will increase proportionally to the increases in clock speed. Of course, overall system performance deals with more than merely the CPU, but tha's another subject entirely and is completely independant of what CPU you use.

-Raystonn

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

 

[Guest]

"...because who I am has nothing to do with my arguments." very cool raystonn, I admire you for that. We always try to form a mental picture of what the other posters on the board look like and what jobs they have or how much money they make but that is never important. It's a sad, sad thing when someone has a great idea but is thrown down because they are not in a certain class or group

 

[Guest]

note to self: "giga=billion" I new that. Duh....
I like that GIPS thing...LOL would that be a gazillion operations per second?

Life is hard...Live with it.

 

[Guest]

It couldn't be GIPS. Millions of Operations per Second=MIPS
Giga Operations per second=GIPS? No, BILLIONS of Operations per Second. BIPS! ROFLOL!!! I love it!

Aklein

Life is hard...Live with it.

 

[Grizely1]

Either one! LOL

---------
I am the first and only one with a 16MB GeForce2 GTS graphics card!

 

[Guest]

>CPU performance will increase proportionally to the
>increases in clock speed."

Add: "for a given cpu (architecture)". Not sure what your position at Intel realy is, but I think your desk is pretty close to the marketing dept. Lets make the world believe Mhz=speed.

>"if a P4 can go up to 10Ghz, wouldn't AMD and Cyrix do so
>to"

>Not with their current designs. They would need a whole
>new core, as Intel has created with the P4.

The current design of the P4 wont do 10 ghz either. Every cpu manufactureres comes out with a completely new core every so many years, and each of these cores undergoes some minor architectural changes every so often. Intel has kept the Pentium pro core alive for a *verÿ* long time (PPro, P2, P3) and has now, finally introduced the new one (P4). AMD has introduced a new core about hmm.. what, 2 yrs ago ? It still has quite a bit of potential for the next year or two, three after what AMD will be ready with a succesor. Its not like AMD is "behind" intel.. they just follow different roadmaps.. Was intel behind AMD when it introduced the Atlhon (vs P3)? Or the K6 (vs pentium) ?Hardly. Just a different roadmap.K6 competed against end-of-line pentiums and early P2's. Atlhon against late P2/3 and early P4's. oh.. why am i typing all this...

Anyway, Raystonn, I really hope your employer pays you to write all these posts.. they are priceless to Intel. Always very subtle, polite, not entirely "incorrect" but always just a tad misleading and/or biased.. fantastic work. If I worked at AMD I would ask some people of marketing to free some time to spend on this forum.. AMDinside and the like (include me if you like) can't come close to your cleverness.

 

[Raystonn]

I said:
"CPU performance will increase proportionally to the increases in clock speed."

You said:
"Add: 'for a given cpu (architecture)'."

That it not really needed. The very first word implies I'm discussing a CPU. I never said performance in general. What I was stating is that a CPU's performance will increase proportionally to the increases in clock speed. Obviously when you increase the clockspeed on a CPU, you end up with two versions of the same CPU. The old one, and the new higher clocked one. They are of course the same CPU. I never meant to imply any differently. A change to a different CPU would mean a change to a different number of IPC. That changes the performance figures completely without regard to clockspeed.

I said:
"Not with their current designs. They would need a whole new core, as Intel has created with the P4."

You said:
"The current design of the P4 wont do 10 ghz either"

I never indicated it would. I said the current P4 core would go to 10GHz. The CPU will of course go through shrinking in the process in the same way the P2 was shrunk to become the P3. There were a few other minor additions, but the core remained the same.

-Raystonn

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