Possible good idea, please have read

[simon]

Ok so people are pretty dumb,

Guy asks a technician in PCWorld to install his hard drive for him,
Techie opens the case and sticks it in , as he is about to put the power cable into the back of the drive the guy says, WAIT, WAIT, techie says what, guy says, can't you catch a computer virus like that,

HEHEHEHEH


If people can follow the manual what to plug in where,
is three extra tubes going to kill them, i really don't think so



Everyone GET A GRIP if you followed this you will see I am just stating facts, I'm not naive tinking they will happen just trying to let people know it could

The mobo idea, I think is actually quite good, as in it would make my life easier and reduce the risk of damage to components, can you deny that while it may not make it mainstream it may get a few people who were thinking about watercooling to actually do it because of less risk.

And if so this could eventually help it go mainstream, in a few years

[jaydee]

Originally posted by simon


Again Jayydee you haven't been reading, why haven't people hit those speeds with liquid N, because as I have stated the chips could be made twice the size if heat wasn't an issue and the could hit twice the speed.
The chips people are LIQUID N cooling aren't the chips that they aren't making, surprisingly enough!!!!!!!

Yes I have read what you stated there about double the size BS and it dosn't mean anything so I disreagarded it as more useless info. The industry Makes CPU's they way they do for a reason.

I thought you may actually have known something about what your are saying but I realize now you don't otherwise you wouldn't be saying it.

And people ARE ignorant (not stupid, just not knowing or whilling to know). I fix computers regularly for people and they don't even know how to install software let alone know what to do with a hose and a pump. Most users don't even know how to take the cover off their case and they do not want to have to take the cover off their case. I install CD drives for people and they have no clue how to do it even though it is the simpleest part on the computer to install. They don't want to mess with a computer they pay all that money for or they would have built it themselves. They like convenience. I am not asuming it I know it as I deal with it all the time.

Water cooling is not mainstream for a reason. If yopu have not figured out why by now there is no reason explaining it further.

You have pretty much ran out of usefull input to argue about so move on.

[jaydee]

Originally posted by simon
can you deny that while it may not make it mainstream it may get a few people who were thinking about watercooling to actually do it because of less risk.

If it where possible maybe, but who cares because it isn't going to happen.

[Brad]

on most athlons oems have an 80mm ducted down onto the retail hsf, so if one dies it doesn't matter

[simon]

I am assumin you know what the criteria for the size of the chip are

They are this(just in case)
Using the smallest transistor technology available amke it as large as possible until the heat output can't be transfered by a stack heatsink


But as I keep saying, this heat output they stop at could be way higher if watercooling was used to transfer heat. If the current chips were any larger a stock heatsink would not take the heat away and it would burn out. If the stock heatsink was instead stock watercooling, well you get the drift.

And by the way the way that no ones grandma can program the VCR but everyones grandkids can put a computer together(no not the 3 year old but the 16 year old can). The people that will be buying the better computers from now on will be able to handle it, the people that can't didn't even have computers at school, your talking about these guys, the guys that buy the high end computers know what they are doing. I didn't say everyone would get watercooling, not everyone has a good heatsink now, some people don't buy the fastest computer.

[Leuf]

If watercooling is reliable enough to be mainstream, then why do you need to move the socket? If you have to move it, aren't you pretty much telling people it isn't reliable? How are you going to market this as a feature when the average consumer doesn't even know what "atx" is anyway?

I have a pentium 60 in the house that's been running for 8 years and a 486 that's getting close to a decade with no maintenance other than blowing the dust out a couple times. The stock coolers on an athlon will do the same thing. Watercooling can never do that.

[jaydee]

[quote]Originally posted by simon

I am assumin you know what the criteria for the size of the chip are

They are this(just in case)
Using the smallest transistor technology available amke it as large as possible until the heat output can't be transfered by a stack heatsink

i am not doubting your technical logic here.

But as I keep saying, this heat output they stop at could be way higher if watercooling was used to transfer heat. If the current chips were any larger a stock heatsink would not take the heat away and it would burn out. If the stock heatsink was instead stock watercooling, well you get the drift.

Yes, but no one wants to mess with water cooling. If you make a CPU run hotter then you have to make EVERYONE even large bussiness use water cooling and pay to maintain it because the CPU requires it. So the OEM market WILL be affected as they would have to use water cooling in all their computers which is not practical for reasons already stated

And by the way the way that no ones grandma can program the VCR but everyones grandkids can put a computer together(no not the 3 year old but the 16 year old can). The people that will be buying the better computers from now on will be able to handle it, the people that can't didn't even have computers at school, your talking about these guys, the guys that buy the high end computers know what they are doing. I didn't say everyone would get watercooling, not everyone has a good heatsink now, some people don't buy the fastest computer.

This is already ruled out from ythe above.

[simon]

This is no way meant to be offensive but have you picked up on a large amount of sarcasm mixed in with a lot of truth.

If you doubt my technical knowledge tell me how they really do decide on chip size.

Another major point is many major servers or big companies already do use watercooling, have done for years, and years.
(NOT SARCASM)

My point in the last bit was again, good heatsinks are considered as mainstream, but only the highend users use them, these will be the people who will buy these computers and only these. I don't know a single person without a computer(ok maybe a few), and I know absolutely no one who couldn't set it up.
You say people ask you to do things that are really easy, and they are just ignorant, Did you know how to put a cd rom in without, asking some one, being shown, or readin in maual or on the net? No really thats shocking.

If the socket was at the bottom it would make it easier for manufacturer to make safe as only a samll amount of pcb needs sealing, also it makes condensation a problem of the past and will give people confidence becasue in the event of a leak nothing could go wrong.NOTHING.

As air flow increases as other components produce more heat as well, so it is needed, computers are getting louder, more dust gets into system so maintainance is unavoidable, there is no but or ifs about this. Everyone will have to clean it up.
If socket is at bottom access will be easier as a small opening could be made to allow easy access.

[jaydee]

First of all yes, I did know how to put a CD rom in without reading the direction, same with the rest of the comp. I have learned through trial and error. I never read a manual on how to build a computer but I could write a novel about it.

And this statment:

If the socket was at the bottom it would make it easier for manufacturer to make safe as only a samll amount of pcb needs sealing, also it makes condensation a problem of the past and will give people confidence becasue in the event of a leak nothing could go wrong.NOTHING.

Clearly shows your lack of knowledge about strait water cooling. I am not going to tell you why but everyone else here will know what I am talking about.

Give it up. You have got negative coments for a reason. i havn't seen any sarcasm though.

[TigerOats]

With the advent of quantum based procs not too far off (2 years) a great deal of research is actually being focussed on reliable and small refrigeration based cooling. Some interesting reading can be found at here;
http://forums.overclockers.ws/vb/sho...ight=cryogenic

[Cyph3r]

not that this relates too much entirely... that 5Ghz link... hehe it's from the Inquirer.... Take that as far as u will... too far and u are really a fool... The Inquirer will take everything completely out of proportion if u will let them

Just so it's out there


Cyph3r

[gmat]

You want brand names selling water cooled (or even frozen) comps ? Go to www.cray.com and chill out.
It's been happening for supercomputers since the very beginning. But who use those babies ? Scientists. Big industrial R&D teams. Military R&D teams. They know quite well what is a computer.

Now for the mainstream the story is different. Mr John Doe doesn't even know where the f*king power button is ! Simon, HOW can you expect him knowing how to plug in a pump and tubings, or even a waterblock ????
maybe 99% of PC owners *never* opened their comp and barely know what's inside.
Here at work, i saw ppl staring at a beeping server. Heck, i immediately noticed the CPU fan died out, and this very knowledge had me promoted to 'know-all-computer-freak' !!!
Mainstream comps DO NOT have to be opened. If they can remove fans or moving parts (main cause of breakdowns), they will !!!
And definitely no move to 'mainstream' watercooling.
(also, the industrial price of stock HSF is less than $5. beat that.)

[Brad]

not all 16 year olds can build computers. For two years I worked at the school computer lab, from 15.5 years old to 17.5 years to be exact.

At my school every student had a laptop. We had people say things like:

my hdd is broken (windows wanted to enter safe mode, upon further probing the person had just turned the computer off when it was starting up because she decided not to use it)

how do I play music on my laptop (he wanted to get mp3's off the 'net)

I've broken the internet (someone had given him a cross over ethernet cable not a normal one)

my computer won't start (forgot to charge the battery the night before)

stop sending me porn (just a random emailing list thing, she thought we were doing it)


etc

I could go on like this for pages, all from students aged 12 - 18. Quite simply, people are stupid, regardless of their age.

[futRtrubL]

Re the enquirer thing, did intel actuallly say room temp, all I can tell is that they said 23% less heat, not quite roomtemp unless you have some good cooling.
Tigeroats, quantum computing is not 2 years off, they have found that atoms can be quantumly linked, they have done this with less than a dozen atoms a time, these lasted a very short amount of time, and they are still not sure what they can do with them, they are pretty much useless with sequantial logic. You tend to wreck the setup when you read the answer, large amounts of ram are needed.

Edward

[jaydee]

Originally posted by futRtrubL
Re the enquirer thing, did intel actuallly say room temp, all I can tell is that they said 23% less heat, not quite roomtemp unless you have some good cooling.
Tigeroats, quantum computing is not 2 years off, they have found that atoms can be quantumly linked, they have done this with less than a dozen atoms a time, these lasted a very short amount of time, and they are still not sure what they can do with them, they are pretty much useless with sequantial logic. You tend to wreck the setup when you read the answer, large amounts of ram are needed.

Edward

To that end, the company will design circuits which will use a new design of arithmetic logic unit (ALU) to reduce circuitries, and is now capable of demonstrating a 5GHz chip at .13 micron process technology which can operate at room temperature. The test chip shows active power reduced by 23 per cent, and standby leakage reduced by 3.5 times.

It says POWER is reduced by 23% not temp.

The Inquirer is actually pretty acurate. It isn't the off the magazine rack Inquirer!!! It is a real computer news site.

[futRtrubL]

And I said heat not temp. All power is turned into heat.

Anyway, people were asking about raodmaps, here's AMD's.

[jaydee]

Originally posted by futRtrubL
And I said heat not temp. All power is turned into heat.

Anyway, people were asking about raodmaps, here's AMD's.

Yes power is heat, but 23% less power dosn't mean 23% less heat. It depends on how that power is being used. They said less circuits also so that also add to less heat ect....

[futRtrubL]

In data switching All power is turned into heat.
Well, maybe they increased the pfysical size of the core so that passive cooling without a hsf can keep it at rt. I think this unlikely though, the 16MHz chip in my Palm grows warm with use with the back off.

Edward

[jaydee]

Originally posted by futRtrubL
In data switching All power is turned into heat.
Well, maybe they increased the pfysical size of the core so that passive cooling without a hsf can keep it at rt. I think this unlikely though, the 16MHz chip in my Palm grows warm with use with the back off.

Edward

They showed pics of the chip in links at the bottom of that article. The core of it did look larger. Less circuitry also should mean less stuff to heat up.

It DID say room temp. This is supposed to be a server CPU though and the tech probably will not trickle down to desktops for a long time. But it does show that it is possibe and can and will happen non the less.

Here is the pic:

[TigerOats]

Originally posted by futRtrubL
Re the enquirer thing, did intel actuallly say room temp, all I can tell is that they said 23% less heat, not quite roomtemp unless you have some good cooling.
Tigeroats, quantum computing is not 2 years off, they have found that atoms can be quantumly linked, they have done this with less than a dozen atoms a time, these lasted a very short amount of time, and they are still not sure what they can do with them, they are pretty much useless with sequantial logic. You tend to wreck the setup when you read the answer, large amounts of ram are needed.

Edward

Sorry to burst your bubble but MIT already have one up and running and if you read the NASA link that is the time frame stated.

[resago]

double size != double speed.

you can't just throw a bunch of transistors onto a die and expect it to go faster. you can only push a single transistor so fast untill it gets saturated with charge. to get faster with bigger, you've have to get MUCH more complex, perhaps some switching mechanism to let one set of transistors discharge while another set take over, but I doubt that technology exists, or if it does, than the switching mechanism isn't fast enough.

smaller=cooler
semi-organic transistors=much cooler, less saturation.

[futRtrubL]

A couple of things tiger. That article is out of date, it sates that semiconductors will only reach 1GHz in 2 years. That article was from the Petaflops Architecture WorkShop, 1996. If they stated quantum (I'll get to that later) computing will be available in 2 years they were off a bit.
What they stated in the article was the arcitecture would be feasable for 10 years (I can quote if needs be)so that would be '06. And that the performance they would talk about would be a 5 year in the future equivalent (last year so also a bit off).
The quantum computing they talked about the is not the QUANTUM computing that gets the headlines,this is quantum tunneling, a way off moving things (usually electrons into and out of atoms) with low energy requirements, this has the side effect of increased speed. This is still a purely serial application.
Lastly this article is superconductors, and not even high temp superconductors at that, they talk about cooling with liquid nitrogen (BP of77.35 °K) , and even worce helium (BP of 4.549994 °K). The 5GHz proc at least has some apllication for high end computing (servers down) while this could only be used in super-super-computers.

Edward

[gmat]

I had the luck to put my hands on a cyogenic pump (cost: more than $100 000). Helium powered. I had to mount it and to tune it for a plasma generator. 5°K baby.
Well, if you've got some money (huh lots of it i mean), helium cooling is quite easy. Just your usual compressor / phase change system.
Now i wonder what a Socket A (and the mobo beneath) can do at 5°K... Just crack, due to thermal shock, i guess. Could be fun.
As for super-super computers they use some Fluor/Carbon Hydroxide i heard. Now they're mainly using watercooling, as it's easier to maintain.
[Edit] For the story we tried to put a (living) fly in the enclosing with the cryo pump. There was ultra-void pressure levels (less than 1.10-7mB) and the fly was still structurally intact. But it couldnt fly ! The little bug was moving its wings as hard as it could with no result. So much fun.

[simon]

OK, about the 16 year old thing, I was exagerating slightly but the point was it took me 5 minutes to work out how to put my computer together and I and all the people you are talking about are worried they will do something wrong so don't bother checking for themselves, but they all could.
For years big computer companies have been offering support to people so they don't have to do things for themselves, this is just to keep them paying inflated prices for new computers and parts from the Dells, and compaqs of the world.
Most peoples grandmas even given the info and reading would still know what to do.

Double size =double no. of transistors = double amount of work can be done.
Not every command has to go through every transistor, also the P4 new hyper transport whatever, it doubles the amount of info sent to the cpu so that if the cpu is underutilised then it will push extra commands through, if the cpu is busy it will push it through when it can, with something similar it could work which transistors to use.
It would be like having a dual cpu system but in one chip.

It would not increase size of chip as they would have to add other substances in that would cause extra problems of energy and heat transfer, the whole industry is based around getting as many transistors into a specific amount of space possible.
This is why chips of same type get faster as building quality becomes better, and they get more transistors into the same space.

[futRtrubL]

An MPU (multiple procs on one die) will NOT yield double speed, it will give slightly better speeds than a multiple proc system. But that is not near 2x speed. Hyperthreading will increase utilization of the procs if the codes running utilise different parts of the proc. And only if you have multiple threads going.

Edward