Slashdot Mirror
Google Demands Higher Chip Temps From Intel
JagsLive writes "When purchasing server processors directly from Intel, Google has insisted on a guarantee that the chips can operate at temperatures five degrees centigrade higher than their standard qualification, according to a former Google employee. This allowed the search giant to maintain higher temperatures within its data centers, the ex-employee says, and save millions of dollars each year in cooling costs."
Wouldn't Intel run into physical limitations that simply don't allow chips to run at that low a temperature? I'm surprised Google isn't considering moving some of its data centres to Arctic locations where you get cool temperatures year-round. We've seen reports of appealing places like that on Slashdot before. (Of course, that would just be a short-term fix before we move the Earth to a farther orbit around the sun to avoid suffocating in our own waste heat like the Puppeteers in Niven's Ringworld ).
If you don't have the clout of a Google-sized organization to buy higher-rated chips from Intel, I wonder if you can basically achieve the same thing by underclocking. An underclocked chip will run cooler, but I don't know if it'll run more stably at higher temps, although I think it would.
Does anyone have any experience with doing this?
I think it'd be interesting to see whether the cost savings in power and cooling is offset by the cost of the performance losses.
You see? You see? Your stupid minds! Stupid! Stupid!
When in college, I heated my crappy little schack by putting 150W bulbs in every light. It was like my own little Easy-Bake oven.
Not really.
No matter how cool the chips run they will put out heat. If you have two chips that run at X and use Y watts you will save power if you can run them a little hotter and use less power for cooling.
See my blog http://ilovecookes.blogspot.com/ for light hearted technical information.
Pragmatically, if they can't run cool, then it's more efficient to run them hot than to spend more energy actively cooling them.
If you were blocking sigs, you wouldn't have to read this.
I'm not sure what your getting at - if by doing this they are saving millions in cooling expense, they are certainly using less energy. What is "going green" if it isn't energy conservation? The fact that the conservation comes from less work for their AC units rather than efficient little processors is immaterial.
Don't expect any company to do things because its right - but good companies will find win-win situations where they cut costs and do things to "Go Green".
Overclockers
The amount of energy you need to use to cool that stuff is quite significant. And, in case you haven't realized it, generating cool air also create more warm air, it's just not in the data center. It's usually vented right outside.
If the chips could run hotter, they'd have to use less energy to cool the data center, and generate less waste heat from that very cooling in the first place.
I'm not convinced that what they're asking for isn't a good idea.
Cheers
Lost at C:>. Found at C.
This sounds like a scenario where lawyers are trying to act as engineers. That works about as well as you might expect.
There are these engineering things, amusingly called "Schmoo plots", that map out a chip's operating envelope of voltage versus speed versus temperature. From those an engineer can forsee how hot you can run a chip before its rise and fall time margins start to get marginal.
There is very little Intel can do to stretch thing by another 5 degrees. It's not something that can be imposed by fiat. Intel engineers have already juggled all the variables to come up with the best performance possible. SOMETHING is going to have to give. Either the chips will have to be selected and graded for speed, lowering the overall envelope for the chips everyone else gets, or they'll have to fudge some other parameters, hoping nobody will notice, or worse yet they'll tweak some variable right to the edge of raggedness, resulting in worse reliability down the road.
Lawyers and accountants generally don't know you can't have everything. let's hope the engineers educate them.
That means you'd need to make up for the lack of processing power with additional CPUs, which would mean more CPUs to cool.
"Guarantee us a higher temp CPU or we will switch to AMD...and tell everyone about it."
That's not really how the negotiation goes in this type of situation where there are two major supplier choices (AMD and Intel) and Google is a relatively small customer when compared with Dell, HP, IBM, etc.
In all likelihood, the negotiation is more of a partnership where both parties work together to create value. Google says, "We buy thousands upon thousands of your chips, but we also pay millions of dollars annually to cool them. We'd be willing to pay little premium and lock in more volume if you can help us save money by increasing the temperature inside our data centers." Google has done the math and comes prepared with knowledge of how much those specs can save them and forecast of chip need of the next 12-18 months, and the two work together to create value. For example, Google might offer to allow endorsements (as they did with Dell for their Google appliances) in exchange for favorable pricing and specifications.
The "do this or I'll switch" tactic only works well when there are many suppliers and products are relatively undifferentiated, like SDRAM or fan motors.
Odds are this is being driven by a data-center engineering team, who are looking at the cost savings of running their data center 5 degrees hotter.
You don't get what you don't ask for.
Intel will do exactly as much engineering as necessary to keep their target market up, and no more.
If the market wants chips that operate 5 degrees hotter.. the engineers will do their job and see if it can be done. Intel will charge a premium for this.
That's business.
Most of the power supply systems for my servers, which are HP G3-5 systems of various U sizes, tend to waste more power as temperature goes up.
This has nothing to do with CPU's though. It is the power supplies on the machines. As temperature goes up, efficiency goes down. At around 80 degrees I noticed a significant larger draw on the power supply with my amp meter.
I had a gaming system with two ATI 4870's and the 800 Watt power supply would crash my machine if I did not run the air conditioner and keep the room at 70 degrees after some fairly long Supreme Commander runs.
I noticed that the amperage would go up, and the power output would go down as temperature would go up.
I have not conducted any experiments in a lab setting with this stuff, but from experience, jacking the temperature up usually makes power supplies work harder and makes them less efficient.
-gc
Got Geometrodynamics? Awe, too hard to figure out? Too bad.
Yes, but way I see this is:
Intel isn't arbitrarily going, "man, we could make chips that run ok 5 degrees hotter, but we're gonna piss everyone off by demanding more cooling. Just because we can." Most likely Intel is already doing the best it can, and getting a bunch of chips which vary in how good they are. And they're getting the same bunch of chips regardless of whether Google demands higher temps or not.
Google just gets a cherry-picked bunch, but the average over Intel's production is still the same. Hence everyone else is getting a worse selection. They what remains after Google took the best.
It's a zero-sum game. The total load on the planet is the same. The same total bunch of chips exits Intel's fabs. On the total, no energy was conserved.
So Google's "going green" is at the cost of making everyone else less "green". They can willy-wave about how energy efficient they are, by simply dumping the difference on someone else.
That's not "going green", that's a predatory approach. Your computers could require on the average an extra 0.5W in cooling, so Google can willy-wave that theirs uses 1W less. They just dumped their costs and their "eco burden" to someone else.
It's akin to me willy-waving that I'm so green and produce less garbage than you... by dumping some of my garbage in random other people's garbage bins across the town. Yay, I'm so green now, you all can start worshipping me. Well, no, on the total the same amount of garbage being produced, I just dumped it and the related costs on other people. That's not going green, that's being a predator.
I can see why a business might want to cut their own costs, and not care about yours. That's, after all, the whole "invisible hand" theory. But let me repeat it: on the whole no energy was conserved. They just passed off some of their cooling costs (energy _and_ money) to someone else.
A polar bear is a cartesian bear after a coordinate transform.
And ideally, I'd come home to find Alyson Hannigan oiled up and duct taped to my bed
You know you're pathetic when they're even unwilling in your fantasies.
They are not asking for the chips to be made to produce more heat, they're demanding that Intel guarantee that the chips will still perform, even if operated at a higher than specified max operating environment temperature.
You would be forgiven for thinking it makes more sense to for Google to insist that the chips produce less heat, rather than that they will still operate in extreme temperatures, since the majority of the cooling cost come from dissipating the chip heat from the enclosed space. But hey, it's Google, they do things a bit different.
And ideally, I'd come home to find Alyson Hannigan oiled up and duct taped to my bed
You know you're pathetic when they're even unwilling in your fantasies.
You're assuming everyone prefers them willing.
"Who is the Journal of Quantum Physics going to believe?" --Stephen Hawking
>> So Google's "going green" is at the cost of making everyone else less "green". They can willy-wave about how energy efficient they are, by simply dumping the difference on someone else.
The difference is that Google is going to actively exploit the ability of those hand-picked CPU's to run hotter. Chances are that the users who would have otherwise received those chips would not reap any energy savings from the capabilities.
At a minimum, Google is contributing here by forcing a vendor to differentiate chips that have a capability of running hotter from ones that don't. No matter who uses that capability, it's a benefit to the planet (versus the alternatives at least).
MadCow.
I used to have a sig, but I set it free and it never came back.
Hard disks. In fact, I am typically far more concerned with long-term issues with my data than with the computing itself. Not to mention, the CPU is NOT the only chip that can suffer from heat issues.
There are two issues with higher operating temp.
One is that you get less drive current from your transistors, so you get less performance (which everyone seems to understand), but this is usually a fairly small effect for 5 degree C.
The _big_ deal with 5 degree C would be electromigration in interconnect metal, which goes up very quickly with temperature. So the difference in failure rates might be quite large.
If there was any deal at all, it's likely that the Intel engineers tried to remove some conservatism from their temperature estimates to see if they could squeeze out 5 degrees from the thermal budget, or perhaps information on the workload itself to get Intel to "bless" the higher data center temperature.
What? American businesses like saving money almost as much as they like making it. It's environmentalism is not as a big motivator as profit, at least in the US. Make being efficient profitable long term, and some businesses will do it. Make it profitable short term and businesses will fall over one another to do it.
“Common sense is not so common.” — Voltaire
What they should do is get the processors to run at 100 degrees C. Then they could use the heatsinks to boil water, the steam to turn a turbine and the turbine to generate electricity. It would be somewhat like the turbocharger for your car.
Slashdot Post Checklist
Bad Car Analogy - Check
Spelling Errors - Check
A solution that won't work in real world - check
Reply to a comment on top 10% of page so people read yours - check