IBM Supercomputer Cooled With Hot Water

1sockchuck writes "IBM has deployed an innovative supercomputer cooled by hot water in a Zurich computer lab. The Aquasar supercomputer employs a chip-level liquid cooling system that can use water at temperatures as high as 60 degrees C (140 degrees F), and as a result consumes up to 40 percent less energy than a comparable system using room-level air-cooling. The system also uses waste heat to provide warmth to buildings, reducing Aquasar's carbon footprint even further."

Comments

Sooo

[binarylarry]

You could prepare soup while you supercompute?

Re:Sooo

[bsDaemon]

a soupercomputer?

Re:Sooo

[Yvan256]

No, a souped-up computer.

Re:Sooo

[countertrolling]

Could get messy. The Styrofoam might melt.

Re:Sooo

[snowraver1]

Maybe if you like cold soup. Soup is normally served at around 80 degrees C.

Re:Sooo

[Mitchell314]

More like ramen noodles and pizza. Gotta be thematically appropriate.

Re:Sooo

[jd]

No problem. If the water is coming away from the machine at 80'C, run it directly to the soup kitchen and connect the hot water pipe from the kitchen back to the computer.

Re:Sooo

[EdIII]

I would think it would soup or computer.

Saw a presentation on this last year...

[Saint Aardvark]

I was at LISA '09, and Dr. Bruno Michel (works for IBM, mentioned in the article) made a presentation on this work (or at least very, very similar work). You can see the presentation, or download the MP3, here:

http://www.usenix.org/event/lisa09/tech/tech.html#michel

Interesting talk, and well worth your time.

Re:Saw a presentation on this last year...

[wvmarle]

The idea of using "hot" water (well above room temperature) makes sense to me; it's so easy to lose heat using evaporation that way (cooling tower) and it takes very little extra energy to transport the heat away from the chips. When I hear that cooling of a data centre can take more power than the actual computers I would say there is an enormous power saving to be made by not using inefficient heat pumps like used in traditional aircon systems.

What does surprise me really is why those chips appear to run best at any temperature up to just above room temperature (usually they are rated for temperatures of up to 60 C or so). Room temperature is defined at 25 C, or 298 K. So a chip runs perfectly well at 298 K but is not considered so reliable anymore when it reaches a just 20% higher temperature of 360 K.

I see this as either a total coincidence, or a serious design issue. Maybe any chip-experts here can chip in on that? Why can't silicon/CMOS work reliably when it's at temperatures of some 80-100 C? It's not that it is anywhere near melting temperature. Also taken the proper material the housing of the chips should also have no problem in handling those temperatures.

Re:Saw a presentation on this last year...

[srinathhs]

thats a good one..

Re:Saw a presentation on this last year...

[Anonymous Coward]

Short answer...because they were engineered to that tolerance, and not beyond.

We could design chips that would run well at 80-100 degrees, but they would have to run slower and probably use larger transistors than current generations. The reason why all the chips run well at 60 degrees is because that's a reasonable temperature to be able to keep them at. They would run faster if we could keep them even cooler with practical approaches. You might have heard about "extreme overclockers" that use liquid nitrogen to get a desktop chip to run at 4-5 GHz. If it were practical and cost effective for everyone to cool their computers with liquid nitrogen, then they would all run faster, but you'd have someone on slashdot asking the questions "Why can't silicon/CMOS work reliably at room temperatures?".

Re:Saw a presentation on this last year...

[Anonymous Coward]

Higher temperature means higher electrical noise, slower transistor switching, and greater resistance leading to more power converting to heat, at a certain point this will become a positive feedback loop known as thermal runaway and the chip will convert itself into a pile of goo, like old AMD cpus were liable to do. This temp. is probably around 80+ C for most chips I guess...

Re:Saw a presentation on this last year...

[vegiVamp]

Thermal expansion, among other things. The heat causes the conductors in the chip to expand (as almost any material does), and because they get longer their electrical resistance also rises.

There is a video

[Jarik C-Bol]

There is a video in the article, as well as a diagram that seems to explain how this works. (the long and short of it is, the hot water cools off quickly towards the lower atmospheric temperature (which allows passive coolers), but is cool enough to remove heat from the chips.

first?

Re:There is a video

[Jarik C-Bol]

replying to myself to explain better. the chips need to be run at 85 C, (185F) and the water is about 60C (140F) this means the water is cooler than the chips, and will remove heat from them, and then, as i said above, it dissipates that heat quickly because of its high temperature. (along the lines of that whole "hot water freezes faster" thing)

Re:There is a video

[Sulphur]

The temperature difference is higher, so the heat flow is more.

As for the frozen water: Hot water gets dissolved gas driven off and cold not. Cool both and the one with more solutes (the cold) gets frozen last.

Evidence of this: Hot tap that was not turned off at the mains snorts.

Re:There is a video

[Jarik C-Bol]

thats what i was trying to say, thank you for clarifying.

Re:There is a video

[TechyImmigrant]

>Evidence of this: Hot tap that was not turned off at the mains snorts.

My hot tap is not turned off at the mains and it has never snorted.

Re:There is a video

[Anonymous Coward]

>The temperature difference is higher, so the heat flow is more.

I don't understand. The temperature difference is higher than what? Higher than it would be for even colder water? That doesn't sound right.

The bit about why hot water freezes faster made sense to me.

Re:There is a video

[jd]

I'm much more puzzled by what a hot water tap would snort. There can't be much that would give a lump of metal much of a buzz.

Re:There is a video

[jd]

Hot water doesn't freeze faster. However, water at 80'C will cool to 60'C much faster than water at 60'C will cool to 40'C, given standard atmospheric temperature and pressure for the ambient temperature of the room. The flow of heat from a hot medium to a cooler medium varies non-linearly with temperature. For example, as you approach the same temperature, the flow of heat approaches zero.

(In other words, if they piped through cold water which was heated to room temperature, a passive radiator would be useless.)

There is a drawback with hot water, though. The temperature gradient issue cuts both ways. As the temperature of the water approaches the temperature of the chips, the heat flow from the chips is reduced. Thus, water at 60'C will not draw off as much heat as water at 40'C, if the chips were to run at 80'C. You've got to balance this sort of approach fairly carefully.

I rather like the Cooling Tower approach (evaporative passive cooling). Basically, you blast the water through a nozzle that turns it into a fine mist. You collect the water that actually reaches the reservoir at the bottom and top it off. The drawback of this method is that it is somewhat bulkier than a radiator system. It is also not a closed system and therefore is a bit more expensive to run. On the other hand, evaporative cooling is much more effective than relying on simple heat flows, so you can get away with a lower temperature gradient at the cooling end. This, in turn, means you get a steeper temperature gradient for the chips, which means they're cooled much more effectively and can therefore be driven much harder without loss of reliability.

Re:There is a video

[Przepla]

Hot water doesn't freeze faster.

Sometimes, it does.

Re:There is a video

[Gubbeson]

Wikipedia has got a nice differential equation (Newton's law of cooling) which states that the rate of heating = [stuff]*(surrounding temperature - temperature of object). If the temperature of the object is greater than the surrounding temperature, the rate of heating will be negative (cooling) and the hotter the object is, the more negative the rate will be (faster cooling).

Re:There is a video

[jd]

If evaporation is considered a possible explanation, then the containers aren't sealed. If the containers aren't sealed and have equal volumes of water, then they have unequal masses of water. The specific heat is the amount of heat per unit mass required to raise the temperature by one degree Celsius. In order to prove that hot water can freeze faster, you must have an equal mass. Since radiation is a function of surface area, you must also have an equal volume. Therefore you must have sealed containers and unequal pressures. This is not the experiment described on Wikipedia.

However, it would be an extremely interesting experiment to perform. I'd actually suggest three closed containers (one with cold water, one with hot water and equal pressure, one with hot water and equal mass). I'd also suggest running thermocouples to the center of each container and plotting the temperature over time. This will eliminate all of the unknowns and variables not accounted for in the original experiment or in the Wikipedia write-up.

Regardless of what the results show, they will be quite interesting and useful for students starting on the topic of thermodynamics.

Re:There is a video

[Jarik C-Bol]

the explanation that i was given for hot water freezing faster (and i'll admit that i am not a hydrologist of whatever so work with me here) that hot water froze faster because the water molecules where in a more energetic state, which allowed them to assemble into crystalline structures more readily.
retrospectively, that does sort of sound like bullshit.

so i'll admit, the hot water example was a extremely bad one based on anecdotal evidence. But hey! my grandmother used anecdotal evidence all the time, and she lived to be... well she's still alive.

Re:There is a video

[Trogre]

I was under the impression that hot water can never freeze.

*shrugs*

Re:There is a video

[LongearedBat]

but is cool enough to remove heat from the chips.

Why do that? Cold, soggy chips don't usually taste as good.

Re:There is a video

[jd]

The freezing point of water is altered by pressure. The triple point (where water can be solid, liquid and gas) is fascinating. I've not looked this up in a while, but IIRC there is a pressure/temperature combo where ice sublimes without going through a liquid state at all. With sufficient pressure, of course, anything at any temperature will set solid. You just need a pressure so great that the molecules are locked into a single position.

Re:There is a video

[jd]

Thermodynamics students would benefit from a correct understanding of heat transfer. If the model is faulty, then their calculations will be in error. If the counter-claims are faulty, knowing and understanding why will prevent them from going in messed-up directions. Extracting useful work is only possible (on any quantitative level) if you know what work can be usefully extracted.

Re:There is a video

[timeOday]

If only they could nudge up the operating temperature of the chips from the stated 85C to something over 100C, they could be making steam (instead of just warming water) which consumes a huge amount of energy without a big temperature gradient; evaporative cooling seems like just a step in that direction. (Heat pipes do use phase change; do most of the use alcohol or something with a lower boiling point than water so the chip doesn't have to operate over 100C?)

Re:There is a video

[Anonymous Coward]

Hot water doesn't freeze faster

Who would've thought. Hot water actually does freeze faster:
http://www.phys.ncku.edu.tw/mirrors/physicsfaq/General/hot_water.html

Re:There is a video

[sjames]

Evaporative cooling is getting to be a real problem in some areas due to water shortage. That could make it anywhere from prohibitively expensive to just plain prohibited during the drier months.

Supercomputer ?

[andymar]

6 TeraFlops is not much, it will not be close to even a 500th place at top500.org.
I realize it's new technology, but it is a bit too early to call it a supercomputer.

Re:Supercomputer ?

[WrongSizeGlass]

6 TeraFlops is nothing to sneeze at, but this is just an experimental implementation ... and it only uses 10 liters of water for cooling (a pump ensures a flow rate of roughly 30 liters per minute). I'm sure they can easily handle a larger water system which could cool many servers which would eventually produce several Supercalifragilflops of computing power ;-)

They should put the chips underneath

[countertrolling]

When the pipes start leaking, it won't pee all over the board

Re:They should put the chips underneath

[Smallpond]

IBM is very good at making non-leaking water hoses. I wish Sears would license their technology.

Re:They should put the chips underneath

[Penguinshit]

That would be a computer whiz...

Re:waste heat to provide warmth to buildings

[compro01]

Not completely AMD, but IBM's Roadrunner system (built in 2008) uses AMD chips in conjunction with Cell procs.

Re:In United soviet-States, Hot-water cools you.

[Hylandr]

You had me until you said 'rooves'. :)

- Dan.

Hypothermia in hot water

[Anonymous Coward]

Water is really an effective cooler even at what you might normally think of as quite high temps.

Reminds me that you can die of hypothermia even in tropical waters of 80 degrees if you are unfortunate enough to get trapped in such water for long enough.

Re:Hypothermia in hot water

[aiht]

Just a note for the confused (like me) :
This phenomenon occurs in water "as warm as 82 degrees F - 91 degrees F" - not degrees C.
This makes sense, because it's still a lower temp. than your body requires to survive.
I would certainly be concerned if it was possible to die of hypothermia at 80 deg C.

Moronic buzzwords "carbon footprint"

[oldhack]

when there is a perfectly straight-forward clear phrase: energy efficiency.

Re:In United soviet-States, Hot-water cools you.

[jd]

Sorry, some conditions are beyond any medication known to man. However, the GP may have a bright future in politics.

Next up?

[gyrogeerloose]

"Home heated with ice water."

taking a shower

[TrancePhreak]

"Honey, we ran out of hot water, could you put on some DOOM5 while computing all of pi?"

Re:In United soviet-States, Hot-water cools you.

[PopeRatzo]

Plasma physics practically transmutes radiant heat into actual matter, so that would be interesting if excess heat was carbonized into coal or perhaps into White light for an inductable surface-level cold fusion illuminance and energy collection from the surrounding environment!

What the hell are you talking about?

Are you in any way related to Professor Irwin Corey?
 

In soviet Russia...

[prxp]

...hot water cools you!

Re:In soviet Russia...

[Anarki2004]

Actually, that's happening at IBM...unless they moved their research facility to soviet russia...To stay true to the meme, you would have to say, "in soviet russia, hot water warms you". Or something....

Re:In soviet Russia...

[hbush]

Nope. In Soviet Russia, you warm the hot water.

Re:This is more idiotic "conservation"

[bigtrike]

How is not driving at all unlike unplugging something? Also, blowing up or destroying vehicles is a stupid way to protect anything, especially since it will likely result in a net increase in every kind of pollution.

Lack of cooling water = overheated chips = ?

[mbstone]

If a lot of people take a shower all at once will this cause a network latency?

Re:This is more idiotic "conservation"

[Krahar]

Wouldn't the same argument say that eating healthy food to prolong life is pointless, since it cannot completely stop aging?

Re:This is more idiotic "conservation"

[Jarik C-Bol]

Holy Balls.

aaaaand then the FBI, NSA, and department of homeland security shut down /. for promoting terrorism.

I have an idea...

[grumpyman]

They could attach the system to the bathroom urinals and gain additional heating for the building.

Re:This is more idiotic "conservation"

[pudro]

People are stupid, and it isn't limited to this kind of "conservation". How many times have you heard someone talk about how much money they "saved" by buying something on sale that they would not have paid full price for?

Re:This is more idiotic "conservation"

[Red_Chaos1]

I think you're missing the forest for the trees. If the new computer consumes less energy than the computer it replaces, then it *is* a conservation of energy. Your whole theory revolves around the assumption that "nothing goes offline, it's just one more thing added" and that is faulty. When it comes to racks of servers, replacing old units with new ones with more efficient PSUs and less power hungry components males quite an impact on how much *less* energy is consumed overall. As long as everything we do pushes the older inefficient stuff out to be replaced by the more efficient models, then we continue to reduce our power consumption.

Re:politics?

[Denihil]

no you don't! join the apathy political movement! we do stuff, but mostly we just, you know...get baked and watch mst3k in our parents basements.

underclocking

[JeanBaptiste]

i knew there's be underclocking enthusiests sooner or later

Re:politics?

[jd]

I considered joining the apathy party, but it seemed like too much effort. I prefer the Official Monster Raving Looney Party anyway.

Re:In United soviet-States, Hot-water cools you.

[manicmike66]

I still remember learning that rooves is the plural of roof. It is quite correct - in English, that is, as opposed to American. The latter has over the past century or so made up their own spelling and called it English. See http://wiki.answers.com/Q/What_is_the_plural_form_of_roof_-_roofs_or_rooves and http://answers.yahoo.com/question/index?qid=20081014042809AArplwb I was taught English in Northern Ireland, not Australia.

Re:In United soviet-States, Hot-water cools you.

[Dr Herbert West]

Roofs?

Re:In United soviet-States, Hot-water cools you.

[instarx]

Ok, so he got "rooves" right. The rest is total baloney.

Re:What's next?

[aiht]

Burn-off to preemptively burn an area in a controlled fashion in order to avoid future uncontrolled fires.
E.g. Burning-off in Queensland, Australia

</smartarse>

because of

[Anonymous Coward]

nonlinear, and even exponental change in resistance etc with increase in temperature

quibbles and bits

[nobodie]

yeah, a geeky quibble: In heating (and i guess therefore cooling) system design "hot water" is water just below boiling, thus 200F to 210F. Warm water is 180 to 200F. This is neither, so to call is hot or even warm is a gross exaggeration. (True story: when I built my own heating system some time ago i ran into this confusion because the engineers and spec sheets kept talking about "warm water heating" which sounded to me like bathing temperature. Totally wrong, and it took a while before I ferreted out the information that brought understanding.)

Drying your hair, grilling a burger

[MightyDrunken]

A flight of fancy of mine recently is the idea of using high powered CPUs as heaters. Imagine if every hair dryer and electric grill was part of a massive distributed computing effort. When my girlfriend is drying her hair should could be computing the folding of proteins!

Shame the least likely part is me having a girlfriend. :(