Just water-cooled my dedicated Q6600 (pics)

Joined
Apr 3, 2008
Messages
43
So I've always wanted a quad-core processor, and I've always wanted to try out water-cooling. While water-cooling isn't the best bang-for-the-buck when it comes to a dedicated folding rig, as opposed to a good air cooling heatsink + fan combo, it was an experience I nonetheless wanted to try.

My results are as follows: Q6600 @ 3.9GHz (433x9), currently accruing 5,400 PPD, running 2x SMP clients in native Xubuntu (configured for "normal" memory units = straight Project 2605s), and more heat in my dorm room (Vcore 1.58V)--temps are 37C idle, 58C folding.

I previously was clocked at 3.54GHz (393x9), air-cooled with a triple-fan-modded Tuniq Tower and accruing 5,000 PPD with the same client/OS setup and Project #s, but with less heat in my dorm room.

Honestly, If I could do it again, instead of getting the tuniq tower and then water cooling I'd probably just get a T.R.U.E. (Thermalright Ultra-120 Extreme).

Enough chatter, the pics:

picture003ik2.jpg



picture001xc3.jpg



picture004ph7.jpg



picture002oe3.jpg
 
Nice work, I love WC setups but they sure do take up a lot of room though. But, very nice work mang.

 
All I could say is awesome! Great OC man. :cool:

 
I could boot into Xubuntu at 3.96GHz (440x9), but the system would crash as soon as folding commenced. 433x9 was the fastest I could achieve stable folding. I might have some headroom in terms of upping Vcore, but with summer approaching, and the amount of heat that this thing is already spitting out, I think I'll keep things where they are (for now :).
 
I could boot into Xubuntu at 3.96GHz (440x9), but the system would crash as soon as folding commenced. 433x9 was the fastest I could achieve stable folding.
That's still more than I know anyone else achieved for a stable folding OC. That will probably net you in excess of 5000PPD for one processor alone.
 
Also not mentioned, I lapped my Q6600 IHS. And I keep a fan on my voltage regs heatsink (not pictured).

This is a photo of my E6600 that's waiting for a new home. My Q6600 lapping didn't come out quite as good (second time using the sandpapers...), but you get the picture (although it's rather blurry).

picturepe3.jpg
 
2 radiators! :eek: Wow! Does a quad pump out that much heat? What's the diff between 1 and 2 rads temps wise?
 
2 radiators! :eek: Wow! Does a quad pump out that much heat? What's the diff between 1 and 2 rads temps wise?

Regardless..... that's a real pretty rig there Mr. Folder.... I am impressed:)

Good stuff....../Pacifico and a tamale to you:cool:

 
Aye, dual 3x120 rads is a bit excessive for cooling just a quad, but this is [H] so fuck it. If you have it, why not use it? If you don't have it, buy it and then use it.

 
Very nice looking. Definitely impressive! I've always wanted to try water cooling. I don't think I'll ever be able to justify it though. I've decided that there is no new computer for me until my ASUS PC-DL motherboard gives out. Then I'm getting a quad. A couple years back I would have been all over water cooling, but air cooling has gotten so good lately.

Do you put anything over the setup to protect it from people dropping things on it?
 
I have a single dorm room, so I'm the only one to blame if something gets dropped on it. So far so good. And I agree with you about the "triumph"of air-cooling as of late.
 
BTW, would chilling the water actually improve your OC? I never use WC before but have always been intrigued by the potential of chilled WC.

 
Very nice...I assume it's "just because" since obviously it is a bit of overkill for a dedicated folder. ;)
 
Nice setup, but why two Rads? A single 120.3 is enough to cool two GTX's and a Quad I'm just a bit confused maybe I didnt' read enough :)
 
That's a dynamite boxen err.. desk my man. (two big rads, all right!!!) :p

I can dig where you're comin' from about air coolin' if you had it to do all over again.. IMHO air coolin' is much less expensive, much less hassle and with heat pipe technology and the new quiet low rpm fans it can get almost as good as WC'ing. Plus you can be "dumber than a box of rocks" and still strap a hundred pounds of copper on your poor mobo without warpin' it too badly.

When I say almost, I mean close but no cigar :D, because IMHO there's nothing like getting a first hand look at a up to date, well setup, dyed water goin' thru the tubes water cooled "moe' sheen" (boxen, computer, whatever, the box with the little electrons runnin all over the place) No matter if you're completely new to computers or an old hand IMO it's just beeeauuutiful. The only thing that IMO might look better is a fine young lady in the raw. (so there's no misunderstanding I'm talkin' about unclothed)

Of course this is only my humble, uneducated and maybe incorrect opinion. (nahh...., the incorrect part is streching things too far, a little humor :eek:)

 
Nice setup, but why two Rads? A single 120.3 is enough to cool two GTX's and a Quad I'm just a bit confused maybe I didnt' read enough :)

Nah -- at near 4Ghz a single 3x120 is just barely enough for a Q6600. Add a couple of GTXs and even a 3x120 will start to go nuclear. A 3x120 for the quad, a 3x120 for the GTXs. Far more realistic.
 
I agree that two radiators may be a bit excessive, but since this was somewhat of an experiment, I wanted to make sure I reached the limit of what I thought was the high-end-safe of Vcore for my Q6600. I'll probably use the extra rad towards cooling a second system down the road--I have a spare pump and extra tubing, so really all I need is another CPU block and, well, another Igloo Playmate cooler. :D

My next build will either be a Q9450, DDR2-1066 RAM, MSI neo2-FR P35, on water or another Q6600 but on air with a Thermalright Ultra-120 Extreme. With the water-cooling parts I already have, and a board that permits a high FSB, even with the 8x multiplier I'm thinking I can hit 4.1-4.2 with the 45nm architecture. Or I may just go "budget."
 
BTW, would chilling the water actually improve your OC? I never use WC before but have always been intrigued by the potential of chilled WC.


Chilling the water would lower temps which might lead to more headroom for overclocking. The problem is that you run into condensation issues once the water starts getting too cold. At that point you're starting to cross into the world of extreme cooling (although not the most extreme) instead of just basic watercooling.


 
I've seriously considered putting a chiller in my machine, but I can't find one that isn't crazy expensive... and the associated condensation issues are a problem, of course.
 
(...) currently accruing 5,400 PPD, running 2x SMP clients in native Xubuntu (configured for "normal" memory units = straight Project 2605s), (...)

One thing I may ask you there... I've done pretty much everything I've seen over the internet to try to avoid the 30XX series projects but all I seem to get here is 3065 love, even when I get a 3062 / 3064 WU is a pain relieve for my poor Q6600 running 2x SMP in Kubuntu. What am I doing wrong?

To Know ... removed -advmethods flag, set maximun memory of 512 Mb per client, and so on. Am I missing something here?

BTW sexy setup you have there, already running a TRUE I can't justify WC just for folding. Keeping it as cheap as possible is my main concern, electricity is already too expensive.
 
BTW, would chilling the water actually improve your OC? I never use WC before but have always been intrigued by the potential of chilled WC.


Chilling the water may help increase the OC however you have to be concerned about condensation around the water block if the water temps get too low
 
Get a fan blowing on that heatpipe on the chipset/vregs and you may be able to clock it more ;)
 
One thing I may ask you there... I've done pretty much everything I've seen over the internet to try to avoid the 30XX series projects but all I seem to get here is 3065 love, even when I get a 3062 / 3064 WU is a pain relieve for my poor Q6600 running 2x SMP in Kubuntu. What am I doing wrong?

To Know ... removed -advmethods flag, set maximun memory of 512 Mb per client, and so on. Am I missing something here?
I've tried nearly everything I know and nothing seems to work. I still receive every type of WU. I believe the only person who claims to have consistently avoided the larger WUs was Tigerbiten. He has Linux VMs without the -advmethods flag and received only P2605 WUs.

Chilling the water may help increase the OC however you have to be concerned about condensation around the water block if the water temps get too low
Condensation even folding at 100% CPU utilization 24/7 at very high OC??
 
One thing I may ask you there... I've done pretty much everything I've seen over the internet to try to avoid the 30XX series projects but all I seem to get here is 3065 love, even when I get a 3062 / 3064 WU is a pain relieve for my poor Q6600 running 2x SMP in Kubuntu. What am I doing wrong?

To Know ... removed -advmethods flag, set maximun memory of 512 Mb per client, and so on. Am I missing something here?

There's a question prompt when you run the -configonly glag that is something like this (paraphrased): "Would you like Work Units with large memory requirement? [small/normal/big]"

Note: this is different than the "Memory to indicate" prompt where you type in a number.

The default choice is "normal" so hit ENTER or type 'normal'.

This seems to be the key config option to get 2605s. I've gotten 2605s with normal memory WUs set and with/without -advmethods flag set.
 
Isn't that Igloo cooler called a Playmate Jr. That under age pron will get you a few years! LOL

Now if you can build an enclosure around that table top and place the radiators into the window AC unit It will be a great cooler yet. haha

Nice work!
 
There's a question prompt when you run the -configonly glag that is something like this (paraphrased): "Would you like Work Units with large memory requirement? [small/normal/big]"

Note: this is different than the "Memory to indicate" prompt where you type in a number.

The default choice is "normal" so hit ENTER or type 'normal'.
You could also edit the client.cfg file and change the packet size. It should accomplish the same thing right?
 
I'm at an "old-school" university with no AC in the dorms (though CT is pretty seasonable during most of the school year), but the rads are drawing air from an open window.

Damn you caught me; it's a Playmate Jr.

Update: @ 3.91 (434x9,) stable--turns out I was at 432x9 earlier not 433x9. :rolleyes:
 
I'm going to mess around with it shortly to see if I can run 4x SMP clients at this higher speed and see if I can still make the preferred deadlines.

Right now I'm doing a 2605 in 9m 15s and 9m 20s per 1%, or 15h 30s per WU (avg.) These WUs have a preferred deadline of 3 days. I'll report back...
 
Nah -- at near 4Ghz a single 3x120 is just barely enough for a Q6600. Add a couple of GTXs and even a 3x120 will start to go nuclear. A 3x120 for the quad, a 3x120 for the GTXs. Far more realistic.

From experience my Q6700 @ 3.6Ghz and 2 GTX's = max 55c with a single PA 120.3 and medium flow Yate Loons (28db @ 48cfm). I do have a 1.5gpm water pump and am running pure distilled which = better cooling. I've run 3dmark06 and game with Vanguard and have never breached 55c. Also my cpu water block is lapped, but that is why I asked about the twin rads... I mean 55c with medium flow fans is great, I could only imagine what the temps would be if I went High Flow at 40db and 78cfm.



 
Just tried that thing, I have a spare folder with a copy of the fah6 so I didn't have to stop any console, replaced the "big" with "normal" in the client.cfg file and guess waht... first WU to download was a 3064. So no difference for me at least. I read somewhere that the SMP console defaults to "big" WU and the option in the cfg file makes no difference, it may make sense after all.
 
Just tried that thing, I have a spare folder with a copy of the fah6 so I didn't have to stop any console, replaced the "big" with "normal" in the client.cfg file and guess waht... first WU to download was a 3064. So no difference for me at least. I read somewhere that the SMP console defaults to "big" WU and the option in the cfg file makes no difference, it make make sense after all.
So, then how are some people able to avoid the larger WUs?
 
You might want to be careful with that open loop like that. With the water just sitting in the open its going to evaporate as i'm sure you've already noticed and its going to introduce more contaminants in the loop. So you might watch your temps after a month or so and if they start to creep up you might have to clean out the block.

IMO, the open loop is just asking for trouble, as those rad's can't exactly be opened up for a thorough cleaning once stuff starts building up inside them. Just a word of caution :)
 
Get Petra's PT nuke, dump a bottle in the playmate Jr and it will kill anything... includign the cat or dog if they drink it!

But I agree get that loop closed or at least somewhat sealed up, lid and duct tape might help.

 
Back
Top