First Retail Water-Cooled DDR2 Memory Tested

Twistedmelon writes "We've all heard of water cooling for processors and even graphics processors, in today's high end PCs. However, a water cooled memory module is something that hasn't been done until now. OCZ Technology recently announced their line of Flex XLC Water-Cooled RAM, with its integrated heat-spreaders that can be connected to any standard water cooling system. The memory operates much cooler under load with tight timings at DDR2-800 speeds. For those with water-cooling setups, these DIMMs could easily be tapped into an existing system allowing for quiet and robust cooling for your system memory as well."

Comments

zap...

[mastershake_phd]

Ive heard of water cooling CPUs, GPUs, and even the Northbridge, but never RAM. Still I guess they are getting hotter too. The only thing I got against water cooling is it uses water, no thanks. Though I would consider using a non-conductive fluid. There is this stuff called fluorinert made for just such an application. http://en.wikipedia.org/wiki/Fluorinert Its extremely expensive though. Ive heard Mineral oil works, then if you get a leak you just get a mess, instead of a fried PC.

Re:zap...

[NeoThermic]

You never fill a water cooling loop with normal water anyway. The two main reasons why is that first it's conductive, and second, it is impure, meaning you'll get problems later down the line with scum forming on the pipes and on the insides of coolers. Instead, the suggested water to use is deionized water, which is non-conductive* and doesn't suffer as much from scum forming (although many still like to use an anti-algae solution to combat the scum that forms.

NeoThermic

(* ok, it still conducts, but it has a higher resistance, and in computers there's few items that'll make deionized water conduct if it leaks. Much safer than normal water)

Re:zap...

[moonbender]

You haven't heard of water cooling RAM because it's a waste of resources, even more so than water cooling other components. RAM doesn't use a whole lot of power and consequently doesn't generate a lot of heat, a quick Google says about 10W. That's comparable to hard drives - the difference being that RAM doesn't really mind running at 50 to 80C, while HDDs do. DDR2 SDRAM doesn't need special active cooling, a somewhat ventilated case is easily enough.

Re:zap...

[CreatureComfort]

Except that what really makes water start to conduct is the impurities dissolved in it. Have you looked inside you computer lately? I've got industrial grade air scrubbers running in my house (roommate with allergies) and I still get dust buildup inside the case. As soon as that deionized water hits that dust, I'm sure the resulting mud would be conductive enough to be devastating.

On the other hand, we are getting closer to the point where everything inside the case that needs cooling could be hooked up to the water system. If we can add power supply and hard drives, then it might be possible to hermetically seal the case, and just cool the water.

Re:zap...

[Phreakiture]

It gives new meaning to the term "memory leak", no?

Re:zap...

[drinkypoo]

Instead, the suggested water to use is deionized water, which is non-conductive* and doesn't suffer as much from scum forming

Ordinary distilled water is close enough to nonconductive, until it gets contaminated of course.

Deionized water is significant in that it is not only nonconductive, but it is noncorrosive. Water is corrosive because of ions known as hydronium and hydroxide (although as I am chem-bozo-man, I have no idea which is positive and which is negative.) Water will actually react with itself to form these ions, so DI (deionized) water is no longer DI after it sits for a while, even in a sealed glass container.

Hence, regardless you will be putting some additive in your water system if you don't want corrosion, and you'll be changing the water regardless. Might as well put in distilled water, since the difference between distilled and DI is pretty much insignificant for this use. You could add alcohol, somewhere from 20 to 50 percent to reduce corrosion. You could use a glycol, like in a car's antifreeze. Hell, you could probably use Redline Water Wetter.

Soon enough.

[shirizaki]

We'll need water cooling for the water coolers.

And then someone will get the smart idea to stick his whole tower in the freezer. Then nerds will become buff by moving around all their heavy equipment.

But how much does it really improve things?

[JoeD]

Every time I see something like this, I wonder how much real world improvement you will see.

Sure, there may be a small improvement on a benchmark, but those rarely translate into something that's noticeable to the end user.

Or is it really more about having the shiniest toys?

Re:But how much does it really improve things?

[penp]

To a lot of people, water cooling isn't only about the possible performance increase, it's also about the sound (or lack thereof). Personally, I haven't tried water cooling yet, but I would definitely like to get the sound of a buzzsaw out of my PC.

As far as performance goes, I recently upgraded from RAM that had a CAS latency of 3 (Corsair XMS) to some that had more aggressive timings (OCZ performance ram) with a CAS latency of only 2. They were running at the same speeds (DDR 400 / PC3200), but at the faster timings the improvement was vastly greater than I had expected. After reading up on it some, a difference of 1ns can mean a lot when you're talking in terms of tens of millions of data cycles.

Re:But how much does it really improve things?

[plover]

Truthfully, it's mostly about adding shine to their toys. Once a gamer gets a frame rate of 72Hz on a 1920 x 1200 display, it's time to add all the detail -- reflections in the water, individual leaves on the trees, beads of sweat on the character's forehead, that sort of thing.

There are also the people who do serious work who might notice a boost in productivity: they might be able to render movie frames faster, or compile a project with 10 million lines in an hour instead of 70 minutes. Of course people in those industries are already massively parallel, but if you have a 60 person development team anxiously awaiting the next build, anything you can do to shave off a few minutes could have really quick paybacks.

Re:But how much does it really improve things?

[SharpFang]

These radiators add at least 1.4 inch to your e-penis length.

Re:But how much does it really improve things?

[moonbender]

Quietly cooling a computer is really, really easy these days. Two or three years ago, it was a pain, because CPUs actually generated more heat back then, but more importantly because silent computing is just mainstream these days.

The power supply is probably the hardest thing to get quiet, unless you're going with a passive one. But you can also get an efficient one with a good fan controller, like a Seasonic, for 75 bucks or so. Silent CPU coolers can be had for 25, and while most graphics cards these days are obnoxious by default, you can regulate the fan speed down to sane levels when you're not actually gaming. (I actually have an aftermarket VGA cooler lying around, no point in using it since the stock fan is essentially turned off 90% of the time.) Another component which is difficult to get silent is the hard disk - elastic mounting works wonders, but it's still just quiet, not silent. You need to go for a more expensive 2.5" drive and possibly sound insulation for that.

So um, the point is, from a silent computing point of view, water cooling is pretty much a waste of time. It might still be interesting for the overclockers who (still) need extreme cooling power, in combination with peltiers or something; I wouldn't know, that stuff was never my thing.

Re:But how much does it really improve things?

[plover]

True, I'd never trust our build cluster to a bunch of potentially unstable, home-made overclocked Franken-PCs. And even if I would, I seriously doubt management would agree with me. They'd have one of two stock answers: "we'll add another server to the cluster" or "there's no money in the budget, live with it."

No extra credit points for figuring out which answer we'll get this year.

However, I will say that the recent set of Dell workstations we got in technically use water cooling. The heat sinks use heat pipes to passively transfer the heat from the CPU up to the large copper radiator fins, and the heat pipes most likely use water as their internal cooling fluid.

Re:But how much does it really improve things?

[moonbender]

Not at all. Just replace the stock 80mm fan with a better (thermally controlled) one.

Yes. That's hard. At least compared to just buying cheap, off the shelf quiet components.

As for the rest: the doctrine at Silent PC Review, probably the definitive resource on this kind of stuff, is that 2.5" disks overall are more quiet than 3.5" disks and that current 120mm PSUs are more quiet than 80mm or 2x 80mm PSUs. The current Seasonics (sorry about the price, I'm in Europe and didn't want to underestimage the US prices) are really great, and they're 120mm. The 80mm PSUs had the advantage of usually having a straight airflow, which is more efficient and which also lends itself for easy PSU tunnel solutions giving the PSU its own airflow.

I've mounted my HD on elastic cord - you can't get more vibration isolation than that - and the noise management set to silent, it's still pretty much the only audible component. And it's a fairly silent HD to start with, 3.5" though.

But of course, your mileage may vary.

I love the smell of Slashvertisement in the mornin

[beavis88]

This is has been done for awhile. Yes, perhaps not DDR2 and memory/waterblock made by the same company, but in my mind, that's the least newsworthy part of this whole endeavor. Nothing to see here (except an OCZ ad), move along...

I concur, it's a Computer "ricer" territory

[spineboy]

Marginal improvement in performance, with the possibility of doing serious harm to your system. This stuff also reminds me of the 100 MPG "vortex" air fans that you put into your car to improve it's mileage, or guys buying big rims for their cars.

Beg pardon?

[Anonymous Coward]

I water-cooled for years, CPU and northbridge, and I can tell you that water-cooling RAM has been done for several years now using home made water blocks. This MAY be the first retail block from a memory company but there is certainly nothing groundbreaking about that.
Water-cooling RAM has always struck me as a lot of work for little to any performance return. Plus it's one more thing to go wrong. I never lost a component in 4 years of doing this but it was such a pain to install and maintain. I can only imagine the headaches involved in plumbing up RAM modules too.

Re:Comparing e-penises

[Aladrin]

Sounds like someone's a little jealous.

Are they hurting you in any way? No.

Are they driving the industry that makes computer products better and better? Yes.

Does having better products every year help you? Yes.

Seriously, just let them have their fun. Most of us can't afford to spend the mega-bucks on things that don't matter, but these people can, and they enjoy it. Let them have their fun while they inadvertently make the world a better place.

Besides, I'm sure you have some hobby that most people think it pointless, too.

Not so

[Gazzonyx]

Clean (Distilled), water is an insulator. In fact, toms hardware inmersed a computer in water in their article Strip Out The Fans, Add 8 Gallons of Cooking Oil. Just used distilled water, and not tap water - as you shouldn't be using tap water, anyways since it eats away at the parts of a water cooling system.
As long as you don't have free electrons, you won't be passing current.

Re:Not so

[modecx]

Distilled water in a cooling system is no better than tap water. Sure, you won't have an amount of ions that will be signifigant to conduct electricity...for the first day, that is. The particles are still going to come off the metals that compose your system, and then you're on the path to being hosed by galvanic corossion.

The key is to not using dissimalar alloys in your system. An aluminum block and a copper radaitor are going to cause problems, unless you use some of the products out there which combat. That's the real key. Pure water is even more corrosive than tap water. Ideally, you want your alloys to be as close as possible, simply for the fact there will be little electrochemical potential.

Not so, maybe...

[Gazzonyx]

Yeah, in my original I wrote that I thought that it had to be non-polar, but I wasn't so sure about it, so I took it out. I'm fairly sure you're right. I think it takes a LOT of electricity to break a double covalent bond, and still quite a bit to break a covalent bond. So, those electrons are fairly well 'glued' in their orbits... But when you take into consideration that the water will be absorbing heat in the presence of something like 2.5 amps (I'm assuming 300 watts @ 120 VDC - I'm sure this is oversimplified and wrong), I'm sure some change takes place. It's been a good 3 years since highschool (or any) chem., and I didn't take AP. Anyone got an authorative answer to this?