Vapochilled Pentium 4 System At 3.3GHz
SpinnerBait writes "Overclocking the Personal Computer has gotten considerably more elegant over
the past few years and there is now an entire industry dedicated to it.
One of the latest innovations is super cooling processors down to sub zero
temperatures with standard vapor phase refrigeration, in an effort to allow clock speeds to crank far beyond
manufacturer specifications.
This article takes a look at the Asetek Vapochill, a Vapor Phase Refrigerated PC
Case, that chilled a 2.8GHz Pentium 4 down to -7C and allows it to run
stable in a workstation environment at 3.3GHz and beyond."
At such high speeds do you really want to spend heaps in order to go faster? I thought the general feeling was that people aren't finding much need for much faster processors. Like most /. readers I haven't read the article yet however I have seen these cases advertised and they cost a bomb. Without doing a price check it might almost be cheaper to buy several lower spec pcs if you want the overall power (say for the seti programme or cancer curing stuff).
Just my 0.02c What do you think?
My blog [.net, rants, general IT]
Is the gain really worth it anymore? I still have an old 1ghz laptop. I use an ancient 333mhz desktop. My server is an extinct 133mhz. I'm all for the "I'll do it because I CAN" attitude, but wasn't overclocking originally for serious benifit? Like 100 - 133mhz? Thats a 33% increase. 2.8 - 3.3 is only about 2%.
You also are instructed to fill all the pin holes in the motherboard socket with thermal grease as well.
What a mess. Just don't try this with arctic silver.
I've had enough abrasive sigs. Kittens are cute and fuzzy.
1. expensive motherboard...
2. expensive CPU
3. moisture on both
No thanks... Interesting, but I don't have enough free spending money to attempt this with such a risk.
A coworker and I, where discussing overclocking and case mods and toying with building a PC in one of those compact refridgerators, maybe with a little modification to the smaller freezer element. It's probably been done, tho. :-)
A feeling of having made the same mistake before: Deja Foobar
It's much more than that. Do your math right.
I still agree that it's not worth the effort, though.
Less than $500 for the kit so long as you don't need to beef up your PSU, and I believe most OC'ers will have a fairly decent one to begin with.
I remember a couple of years ago someone in the comp.sys.ibm.pc.hardware.chips newsgroup paid $5000 to get a Kryotech system that allowed an Athlon 7-800MHz to be overclocked to run at 1GHz. Madness!
I've got a fever and the only prescription is more COBOL.
Isn't there problems with condensation on occastion?
All that money and trouble for a measly 500MHz. Sometimes I question people's sanity.
Of course, and I highly suspect it, I may be talking out of my ass. -oqti
at this price point, shouldn't people be thinking about adding additional processors, instead of overclocking 1 processor?
That seems like the better path to follow from a price, performance, and stability standpoint.
Need to get away?
Adirondack Vacations
For those who use a real method of telling the tempreture, the tempreture is -7C/266k and NOT 17.9!
"Titanium" case with cool LED blinky-ness... now if they could get the thing to have a built-in fog machine, I'd be happy.
Anyone know where I could buy a cheap a gas turbine engine?
I've always wanted to know how just far you can get with basic air cooling.
"I have opinions of my own, strong opinions, but I don't always agree with them." -- George H. W. Bush
As lovely as it would be to have a compressor running in the room while you're using your PC, isn't this only going to give you only limted benefit - besides the wank factor of runing at 3GHz or whatever? Unless you're overclocking your memory as well, your P4's going to run out of gas no matter how fast it's turning over. And I don't see an fridge adaptor for your memory.
Reliable, Great Value Hosting: $7.95/mo 2.4G/120G
virtual + 1, funny.
gone are the days when overclocking was actually an economically viable option to get more performance out of your system for what you pay. these days, overclocking is almost no different from case modding, in that its just for posers to make themselves feel better and have something to show-off.
and is it a good sign that slashdot is continually posting articles pertaining to both case modding and overclocking "breakthroughs"? yes, they overclocked the system to 3.3ghz, but most likely in a few months intel is going to release processors that are just as fast, if not faster. see the pointlessness?
its like posting on article on slashdot about a breakthrough in man-powered vehicles, about how 200 people got together to push a car to more than 200km/h (i refuse to use miles/hour). yea sure, that was fast but is it a breakthrough? currently, the approach to overclocking processors is to up the voltage so the processor can function at a higher clock speed reliably, and then find some way to cool it down sufficiently so it doens't overheat. so they got it colder this time, and faster, via the exact same approach thousands of overclockers have been using all the while.
is this really news-worthy on slashdot?
The general feeling is indeed that there isnt much need for faster processors. ;))
This article, however, has nothing to do with the general computer user. Its about having the fastest available chip on the market and running it faster than everyone else. While there isnt really a practical use for this, some people enjoy it (myself included. I tend to stick more to the "budget" end of things though
I could use a car analogy here, but thats been so overdone I'll spare you.
Phase - Change cooling systems such as the one used here typically only cost around $700. Thats less than twice the cost of a high end graphics card and it can be used to supercool any processors released using the same socket system. In short, its not such a huge/useless investment as you make out.
This is my first post on slashdot, so apologies for any mistakes.
~Laurence Wood
the idea is simple. on top of the chip one places a vertical tube with the same crossection as the chip. The tub is filled with alcohol or propane or freon or other low boiling point liquid. The sides to the tall(!) tube are lines with air-cooled heatsinks.
when the liquid boils then the (VERY LARGE) heat of varorization is extracted from the liquid. the expelled gas molecule rapidly transferes its energy to other gas molecules and then distibutes that over then entire face of the heatpipe which condences the gas back to liquid.
the processor can never warmer than the boilingpoint of the liquid. the average cooling capacity is determeined by the requirment that the cooling rate of the heatsinks equal the heat input rate on average. One of the nice things about this as opposed to a fan or refregeration system is that although the average heat load is the same, the peak heat load can be as high as you want. the liquid has almost infinite reserve cooling capacity up until it boils dry. Thus the temperature of the processor fluctuates less than any fan cooled or refrigerated system.
So what is the heat load capacity. It should be the significantly larger than any refrigerated system with the same area of heat sink!!!
Some drink at the fountain of knowledge. Others just gargle.
Seriously, I really don't give a shit about overclocking. I don't want to block out all hardware news though.
I often wonder how many slashdotters overclock cpu's of 500+, because I wouldn't want to blow my money buying a new processor. Therefore, I usually skip these articles. Could we have a poll on the lines of:
4 99mhz /harmonica/large HORN/ over my processor. ...or whatever is realistic.
My CPU is overclocked:
0-49mhz
50-99mhz
100-199mhz
200-299mhz
300-
500-999mhz
1GHZ+
CowboyNeal stands in the corner blowing/playing a
"You know you don't act like a scientist, you're more like a game show host." Dana Barret
Just put a tiny "Type-R" sticker on it, and be done with it.
Sheeesh.
at this price point, shouldn't people be thinking about adding additional processors, instead of overclocking 1 processor? That seems like the better path to follow from a price, performance, and stability standpoint.
Yeah, if you want to run a 500W power supply and push that much heat (and noise) out of your case into the room.
Don't get me wrong, I think we're heading in the direction of multiprocessor systems, what with all the recent talk of hyperthreading, but we're not quite there yet.
Overclocking isnt supposed to be a "useful" thing to do anymore. Chip speeds are increasing fast enough and are frankly fast enough already, that there is no actual NEED to do it. Is there any NEED for people to supercharge their cars? Do they do it anyway?
It's basically the same thing. These are just people trying to push what they've got as far as they can. The point isnt the cost, or even really the gain. They're just trying to see what they CAN do, and how to do it. You may think this stuff is worthless now, but wait and see how long it is before these radically overclocking cooling techniques become commonplace in home PCs (once scaled down a bit).
Can someone explain to me why it makes more sense to do this with pentiums for piddly improvements in performance.
Some drink at the fountain of knowledge. Others just gargle.
Cons:
Still somewhat pricey but cheaper than similar competitive solutions
Motherboard tray can be hard to work with
Retail channel for product is still somewhat limited
Shouldn't one of the cons be the electricity needed to keep it cool?
Live web cams
This kind of balls-to-the-wall turbo-charging is not only un-necessary, it's wasteful too.
Most PC's are pretty power hungry as it is, without introducing a whole new load of cooling equipment. Although it appears to make the processor perform more efficiently, actually it makes the whole box a whole lot less efficient in terms of power consumption.
Much more encouraging is the recent trend to making silent PC's. These tend to be pretty energy efficient as well as nicer to have about the place.
All things in moderation; including moderation
Can we get a consensus or perhaps a poll that shows: 1. A minority of slashdotters are case modding and/or overclocking. 2. They/we would all be happier if they could exchange their thoughts in the privacy of their own little 'mods' section. Or at least fast-track them directly into hardware. These stories are largely redundant and lead inevitably to the same bunch of threads. Furthermore, eaking a 20% to %30 speed improvement from your P4, only to be outpaced by next month's offering from Intel -- to incur all that wasted-time, money, and risk, makes about as much sense as pinning neon tubes to your box. I'm going back under my bridge now.
.... move to Iceland and get sub-zero vapochilled cooling for free.
Is there room for it to keep beer nice and cold? For about $500, it should.
Don't transistors (or whatever is inside processors) go faster/work better when they are hot?
I read in the headline that they got it down to -7C. At that temperature, shouldn't the nature of slower cooled transistors kick in and fundamentally (I've wanted to use that word all week) make the processor slow down? To achieve a balance, one of the many hardware review sites should run benchmarks with the CPU at different temperatures and maybe find out a temperature in the middle (~25C or 30C) or on the extreme hot side before it damages/shuts down (~55C or 60C).
Those temperature figures might be wrong, but you get the idea. Maybe the solution isn't to cool it the absolute maximum but to use or create a control system for this compressor to get the temperature constant no matter how much heat is generated.
I hope I conveyed what I'm trying to say and if I read this post again later today, I'll probably think I had a mental deficiency.
A better overclocking solution is the Prometia from chip.con (whose server seems to be down at the moment) which cools the processor down to -40C rather than Vapochills comparatively tame -20C. Get a list of reviews comparing them here.
C'mon Cowboy Neal, this is a waste of space.
They could have gotten much better preformance if they had used a 2.6 Ghz P4 witha 400mhz (100x4) FSB, and overclocked it. The problem with the 533mhz bus chips is that they are essencialy overclocked straight from the factory, thus making more gains much more difficult. Look on any serious overclocker website, and you will find watercool kits (not even sub-zero cooling) approaching 3.8or 3.9 ghz using 400mhz bus chips
Yawn.
No he is right. Resistence goes up as temperature increases which actually slows down processing as the signals become more noisy. The higher operating temperature may speed up gate transistions but it makes the devices more unreliable and the timing becomes less stable. As the chip has to be synchronised to work you need to keep the temperature low to keep it running. This is why OC'ers cool their systems and see how fast they can push the processor (nudging up the heat) until it becomes unstable and you can't run software any more without it crashing.
Slashdot: where don knuth is an idiot because he cant grasp the awesome power of php
It's obvious, or at least it should be, that this type of refrigerated cooling kit, is for the enthusiast. Someone here hit the nail on the head though. It is very much a kin to the type of some folks are VERY interested in, that are tricking out muscle cars.
:)
This system can indeed give you a fast stable processor speed (I've seen one in action at a show myself) far faster than anything on the market from Intel or AMD. Someone noted, "why not wait until the 3.3G is released by Intel"?... or something like that.
You're missing the point, when Intel does release this chip, you can super cool and overclock it as well and probably hit 4GHz. Not too shabby, if you ask me. Then again, you have to be into this type of thing, so if you're not, don't knock it until you try it.
As far as this Slashdot post being worthy, I think it is. It's an evolving technology that you may find much more commonplace in the future.
Hehe... I do agree though... if you've got to put a frige in your PC, you had best find a spot in it for some beer damn it!
on the other hand, how loud are these things? i'm eyeing a system to use as a home theater deck (new TVs have RGB hookups. mmm counterstrike on a 65" tv at 1280!) -- but being in the main room, it'd have to be near silent. for that pc i may choose silence over temperature.
hmm. decisions, decisions.
Oooh you mean I can get another 15% performance is I install a fridge in my PC? Why don't I just wait a month and buy the next version and save the money, hastle and electricity.
Quick question, if you put 6 refrigerators in a room, will the room be hotter or cooler (whether or not you leave the doors open?)
No, I don't think I will ever change my machine for a steam machine
--
ACid
TEMPERATURE you fucking cunt
These phase change coolers have proven themselves to be somewhat destructive to P4's in the long term though. After a couple of months of use, the chips mysteriously die. Now typically when chips die in overclocking its due to electron migration from rediculous overvolting (and this is more true with .13u chips). However several people have had thier P4's die with only nominal (~1.6 volts, most p4's defaualt to 1.5) overvolting while useing these phase change coolers. Turns out, Intel only specifies these chips for operation to temperatures down to 20 F, and what happens is the supercooling causes the chip to become so brittle, that even slight virbrations can cause damage to the chip, thus killing it.
...and IN SOVIET RUSSIA, beowulf clusters imagine 1, 2, 3 profit!!!! jokes made out of YOU!!!
All my games run.
All my software run. I still have to add a little more RAM but windows boots quickly enough (read line 1 for reason why I'm running windows not linux).
I *could* overclock the chip. In fact my motherboard is made to allow this to be easy. But the question is this... "if it's not slow, why risk making it a coaster?"
All you overclockers, we read a new overclock story every 2 weeks or so. It's not really that cool anymore. In fact, many of us just find you wasteful and silly.
Find a way to build your PC into a car, or a robot or something, then we'll pay attention.
Recognition is about doing something new or at least out-of-the-ordinary
I have my P4 Celernon 2.0 Ghz clocked at 2.9 Ghz with the OEM fan on it, nothing else. 3.3 from a 2.8 is not so great.
I do a lot of VHDL at work, and let me tell you, it sure as shootin is worth it for me. Large FPGAs take forever to synthesize, even on mighty big hardware. The industry (at least where I work :) is moving away from Solaris to Linux & Win 2k solutions. The reason is largely bang for the buck and speed..
So, I have a watercooler on my PC. I upgrade the CPU every couple months. That watercooler always lets me eek out another couple hundred Mhz of speed where I couldn't before, and it does make a difference. 10 minutes here and there does add up.
It isn't as signifigant as it used to be, but it is still definately worthwhile. My machine is also much quieter without the 60mm turbine on top of most CPUs.
..don't panic
If you did this with a p4, it would
achieve consciousness.
For a coupla minutes;
http://totl.net/Eunuch/
Pah, I spit on your puny -7C
and minescule performance increase!
In the free world the media isn't government run; the government is media run.
There is absolutely nothing insightful in this post, or the long string of mod 5 posts prior to it. This just reconfirms what I already know: Moderators look for something they agree with and moderate it up to push their perception of the world. A TRULY insightful post is one that you don't agree with, but can understand and appreciate.
Yes, and nobody will ever need more than 640 kBytes of RAM.
But the greatest Electrical Pioneer of them all was Thomas Edison, who was a
brilliant inventor despite the fact that he had little formal education and
lived in New Jersey. Edison's first major invention in 1877, was the
phonograph, which could soon be found in thousands of American homes, where
it basically sat until 1923, when the record was invented. But Edison's
greatest achievement came in 1879, when he invented the electric company.
Edison's design was a brilliant adaptation of the simple electrical circuit:
the electric company sends electricity through a wire to a customer, then
immediately gets the electricity back through another wire, then (this is
the brilliant part) sends it right back to the customer again.
This means that an electric company can sell a customer the same batch of
electricity thousands of times a day and never get caught, since very few
customers take the time to examine their electricity closely. In fact the
last year any new electricity was generated in the United States was 1937;
the electric companies have been merely re-selling it ever since, which is
why they have so much free time to apply for rate increases.
-- Dave Barry, "What is Electricity?"
- this post brought to you by the Automated Last Post Generator...