New Water-Cooled Hard Drives Coming
CoolHandLuke writes "NEC and Hitachi are teaming up on a liquid cooling system for hard drives. The goal is to cut down on noise levels while providing more efficient cooling. 'Hitachi and NEC are developing the water-cooled hard drive systems for desktop computers mainly to reduce noise levels to 25 decibels, 5 decibels quieter than a whisper. To do this, NEC and Hitachi actually wrap the hard drive in "noise absorbing material and vibration insulation." According to Hitachi and NEC, the cooling cold plate they're planning to use is the most efficient plate ever used for heat conduction, which means they'll be able to cool the hard drives quicker and more efficiently.'"
Nothing for you to see here. Please swim along.
...and be enlightened.
"[Regarding the 'cloud,'] ownership was what made America different than Russia." -- Woz
The biggest factor in keeping them down to acceptable levels is just mounting them on soft rubber grommits.
You also need fans to keep them cool. After the CPU and GPU, the hard drive is the hottest thing in your computer. Especially in drive arrays or servers, they can heat up extremely quickly with sustained usage.
If this works like it sounds, then it will not only quiet the drive, but cool it more efficiency and allow less external cooling (fans), which should quiet things down even more.
I don't know what kind of crack I was on, but I suspect it was decaf.
I sure hope it's a LOT quieter than a whisper. If my machine sounded anything like as loud as a person constantly whispering I think I'd go out of my mind.
A-Bomb
Hard drives are very easy to cool. The water block plates don't fit snuggly against the uneven surface of the hard drives' sides. But it doesn't matter. The whole drive is still cool to the touch, much cooler than decent air cooling can manage. In short, there really isn't a need for "the most efficient plate ever". And it won't do you much good besides, as last I checked, hard drives are not very flat on the sides or bottom.
Acutally, yes, I do.
Its more than noise, however. We don't need a more efficient cooling system, we need a hard drive that uses less power and generates less heat.
The whole path that desktops are going down (except for the occasional exception such as a mac mini) is one of more power, more heat, more fans, more noise.
This is, to my mind, the grossest abuse of Moore's law that can be had. Instead, we should be building smaller and lower powered devices. Perhaps it simply reflects how cheap energy is that we choose to build computers this way.
So now we can build a whole class of hard drives that suck more power from the wall, confident that they won't make as much noise?
Am I the only one who sees the folly of this?
Michael
There is no cryptographic solution to the problem where the intended receiver and the attacker are the same entity.
You must be one of those guys with the ipod at full volume in the subway.
:-)
I'm glad you are making something out of your deafness though
Solid state is the future, but right now it is very slow, small, and expensive. Until they start offerring them at flash chip prices and in a SATA format so you can RAID enough of them together to get some usefull performance numbers stardard HD will still be on top.
Same way you abuse anything else - quantum physics (nuclear power or bombs), chemistry (medicines or nerve gas), arable land (corn for tortillas or petrol additive).
The abuse that I see here is that we should be smart enough to not abuse our knowledge and resources for short term gains at a long term cost.
What we do with the evolution in technology is our choice, certainly.
It was not so long ago that people used to laugh at why you would need a 200W power supply for a computer. Now it seems that 500 watts are common enough, and some are going significantly higher than this.
There is no law of physics that demanded this increase in power consumption. It was a choice by manufacturers and consumers.
There are certainly some times when it makes sense to throw the power at the circuitry, but for the most part its just wasted time. To my mind the ideal computer would run at close to 100% CPU utilisation all the time, but the whole system would reduce its power and speed to match the load requirements. Likewise, standby power should be very close to zero - we do this for laptops so much better than for many desktops.
I guess its my personal ethos showing here. Nothing more, nothing less.
Anyway, I hope that explains my position on why I think its an abuse. Energy is cheap, but it may not be for too much longer.
The world really doesn't need a hard drive that sucks more power quietly, at least not for most computers.
Hopefully in a few short years flash drives will overtake hard drives and everyone wins.
Michael
There is no cryptographic solution to the problem where the intended receiver and the attacker are the same entity.
If that is too much to ask for then they just need to set up rows of SATA/USB ports that just let me plug 10-15 memeory sticks straight into the mother board so that a standard RAID solution can be used.
It was not so long ago that people used to laugh at why you would need a 200W power supply for a computer. Now it seems that 500 watts are common enough, and some are going significantly higher than this.
They're nowhere near as common as enthusiast sites would have you believe, and even in most machines that have them, they're not using anything close to 500W of juice.
The average PC is a low-end desktop that probably barely even peaks at much over 100W of power draw.
There is no law of physics that demanded this increase in power consumption.
Sure there is. It's the cost of greater computing power, within the limitations of current technology.
Hopefully in a few short years flash drives will overtake hard drives and everyone wins.
Flash drives aren't going to replace hard disks in the near future as they are highly unlikely to be able to come close to the same size/cost ratio of hard disks.
This plays out every time exactly the same way and yet people fall for it every time. Just like SLi, water cooling is unnecessary. It is and has been used for bleeding edge performance *before* the technology is really at that level and certainly before the software is. If you wait 6month's to a year the hardware catches up, doesn;t need water cooling and around that time software begins to emerge that utilizes the new technology. So for 6 months you gain the ability to say oooh look at my watercooled setup that does nothing except maybe pump out high synthetic benchmarks since nothing utilizes it yet. Same with SLi.
No developer is going to produce for a market that may be 1% of the total market. Once the technology reaches mainstream use (which watercooling and SLi never will) they then begin to utilize it. This has gone on since the days of mainframes, and continues in cycles right up until today... when will people learn?
http://teasphere.wordpress.com - A little spot of tea
Right now, my main PC has three fans devoted just to keeping my drives cooled.
My Tivo has long had the lid off and there's a special fan clamped next to it dedicated just to cooling the drive.
That's four fans worth of noise just for drive cooling, far and away MUCH louder than any other noise in this room.
If there's a way to keep them cold and quiet, I am in.
Sig for hire.
Personally I've moved to using iSCSI on my desktops. Every single one is booted off PXE and mounts its disks over the network (with very close to native performance; gigabit copes well with iSCSI, and the memory in the iSCSI target machines works nicely as cache). Blessed silence ensues, with care for CPU and PSU fans, the desktops become close to inaudible.
The server cabinet is slightly more noisy, but with care taken to soundproofing and with sound-absorbing vent channels and the disks mounted on vibration reducing material, it doesnt sound more than modern fridge.
Adding yet another cooling bus to the desktop sounds like a supremely unpalatable idea. It's much easier and much more reliable to move data over the network than it is to move water around in a computer.