Just water-cooled my dedicated Q6600 (pics)

I think that this 30xx series WU are preferably sent to Quads and Dual core systems are assigned the all wanted 26xx series. A Quad after all have "double" the possessor power, that's why people use two VM, as every one of them show's a system with a dual prossessor.

Don't take my word for granted on this.
 
I had to drop the overclock to 3.9GHZ (433x9) to reach stability with 4x SMP clients, but as a result I'm utilizing some of those unused processor cycles from running *just* 2x SMP client and netted 58 PPD from the addition of 2 more SMP client , My PPD on the Q6600 went from 5460 PPD to 5508 PPD.

Each SMP instance is averaging 18m 24s per 1% and ETA is 1d 6hr 30m for 1 2605 WU.

I'll have to monitor for more long-term stability, so these are just preliminary result (3% complete per SMP client), and to make sure I continue to only get 2605s :cool:

EDIT: Not stable at 433x9 and 4x SMP client. Dropping "more clients is more" theory, and testing a 436x9 overclock with 2x SMP clients...
 
I think it may be a couple of things:
- Memory setting (anything above 512 fetches big work units)
- BigPackets=yes

There may be other settings, but these are the ones that I have seen that affect ability (or not) to get a larger WU

 
Good idea about closing the loop (i.e. topping the reservoir/playmate). I have 2 drops of Dawn liquid detergent in the loop to break down the surface tension (i.e. "poor-man" water wetter) to make the distilled water more heat conductive, since surface tension blocks heat. This additive should help things from getting cloudy. but I still need to cover the thing to prevent the dust-type build-ups.
 
bipolar-folder said:
Good idea about closing the loop (i.e. topping the reservoir/playmate). I have 2 drops of Dawn liquid detergent in the loop to break down the surface tension (i.e. "poor-man" water wetter) to make the distilled water more heat conductive, since surface tension blocks heat. This additive should help things from getting cloudy. but I still need to cover the thing to prevent the dust-type build-ups.
Click to expand...

Yeah some kind of cover would work fine although that makes it tough to monitor your water level. You might rig some kind of float with a straw or something out the top so you can see how much water is left. Maybe mark the straw with a pen with full and empty lines.
 
Get an EK or swiftech reservoir. You don't want the system open like that, stuff will build up inside the rads and the block.
 
The pump I use is submerged, so that limits the type of enclosures I can get for it. My new idea is to, using my Swiss army knife min-saw, see if I can cut some peninsula shaped "holes" in the playmate jr.'s drop-top lid, just big enough to fit the tubing. Easier said then done...
 
@ bipolar-folder, my man, that's one fine lookin' rig you've got. :D

Don't sweat any criticism about havin' two 3x120 rads, from my experience with having 2x WC'ed 24/7 folding duel core E6600's (OC'ed to 2.8 GHz and 3.0 GHz) and a few more learning experiences IMO it's better to have too much cooling capacity than not enough. (who knows, if I'm lucky enough to get a quad I might go with a 3x120 rad or two ;)) I have to agree with a previous poster about you using a "high end" over clocked quad and a couple of GTX's, it's much better to be safe than sorry, who knows how hot dem chips get :(

I also have a "on the floor, in a cardboard box" jury rigged external WC'ing system. It has the parts from a Nautilus 500 external WC'ing rig (I gots it by accident on fleabay) and whether my paranoid behavior of being upset about the box coming with the pump (MCP350) upside down, having a aluminum radiator or only having a single 1x120 rad (seemed insufficient to me) is really nobody's business but mine. I have turned the pump right side up, replaced the aluminum rad with a 2x120 MCR220 copper inside rad and now I'm semi satisfied.(it no longer all fits in the original Nautilus 500 plastic case) I do like the WB for the cpu though, not as good IMHO as a Apogee GT, but pretty nice. I like the "quick disconnects also, I've never had them on any WC'ing rig before, they seem handy for a external WC'er) :p

This is not any form of criticism, but it looks like you have an open type reservoir (your Igloo cooler) with a pump submerged in it (I've never used a submerged pump, my pumps have always been in line, out of the coolant). I think you'd be better off with some type of enclosed rez with, as stated previouly, something like PT Nuke in it. Why not just use a "jug" or ghetto rig something with an air tight cap or top? It seems to me enclosed WC'ing loops are hard enough to keep clean and not growing bugs or plants in, let alone an open loop. :eek: Just use any type of plastic jug or container, I've read about some people using pickle jars)

OBTW, I've been "lapping" cpus since I lapped my PD 805. I lap everything now, my HSF or WB's contact area (sometimes you can't lap some WB's, HSF's , always "razor" check first), my mobo chip HS's (NB, SB, any onboard vid or lan, etc), my cpu HS spreader, (voids the warrenty, I really don't care :eek: ) and anything else that appears to benefit from lapping :)

Sometimes lapping helps as much as 5c or so degrees in my own experiences, I've read where some people get better results :rolleyes: and sometimes it doesn't help or makes things worse, you gots to be careful. (in the case of convex or concave on purpose HSF's or "bowed WB's", cpu HS spreaders or anything else)

 
You must log in or register to reply here.
Back
Top