Single-Loop Limitations? Have I reached it? CPU, GPU, MB, GUP?

C

crewxp

Limp Gawd
Joined
Jul 27, 2009
Messages
393
Hey, I'm running a single loop, a MCP655, and a Triple Rad. I have one video card already (4870) water cooled, but am adding a second 4870 in crossfire mode.

Would it be alright if I add that to the loop? Or should I start looking at a double loop if I'm going to do that.

So if I do this, it will be:
--WATERCOOLED
i7 920 CPU
HD4870
HD4870 CrossFired
Rampage II Extreme Chipset
-POSSIBLE (Is this a no go?)
Hard Drive (Mine get hot as heck)
 
What ? It will take a lot more than that to melt the tubing or the rad core. Expect your loop temps to rise some which can be offset by more air through the rad or another rad or another loop with a rad.

My point is with sufficient airflow any dual rad can dissipate a 1kW of heat easily, you have a triple. After adding the additional video card your CPU and loop temps may not qualify for e-peen bragging rights but that is your call depending on how much money time and effort you want to put into the system.

Benchmark you system now. Temps etc. Add the card. See what happens. Adding another rad to current loop is easy. Adding another entire 2nd loop with its own pump etc. is not trivial but still, its all just about finding where to mount stuff and running some hoses. Its not like you cant change your mind later and do something differently.
 
What kind of rad. What vcore/speed on the cpu. What are your current cpu temps.

As far as a triple rad dissipating 1kw of heat, i cant agree with that. I'm running tri-sli on a swiftech 320 rad and using push/pull fan setup, the water temps stay pretty darn warm compared to my cpu loop with its own 320.
 
As far as a triple rad dissipating 1kw of heat, i cant agree with that. I'm running tri-sli on a swiftech 320 rad and using push/pull fan setup, the water temps stay pretty darn warm compared to my cpu loop with its own 320.[/QUOTE]

\ vicious sarcasm warning !!!
What a surprise your loop with 3 graphics cards each putting out as much wattage as your CPU is warmer than your CPU loop. And so that the scientific testing term "pretty damn warm" gets its proper due I am submitting your findings to the National 6th grade Science Fair on your behalf. :p :D

Good testing is difficult to find and actually I base my statement on an automotive heater core much much smaller than a 3x120 in a 1951 Olds Delta 88 Coupe producing so much heat it would blister your feet through your shoes while driving in winter snowstorms. (Confirmed by personal testing/experience.) Rads dissipate heat until they melt and you cannot guage rad performance on your setup, my setup, or any setup unless subjected to standardized methodology and even then assumptions are made that may or may not apply. For example; How loud, how much CFM can the user stand for the fan(s), it is subjective. Same with acceptable temps.

I hesitate to post the link below because everyone that tests these things uses a different method or is focusing on a different aspect of rad performance. What I want to direct your attention to is the table in about the middle of the article with the orange header, where with a single 2300 RPM fan (high speed but not ridiculous) no shroud, and a 15C air to temp delta (performance nothing to brag about but acceptable ) a SINGLE 120mm rad dissipates over 400W. With a jet engine 2800 RPM fan over 450W. Consider what the dissipation would be if the deltaT where allowed to be 20C, meaning that in a room with an ambient temp of say 22C the loop temp would rise to 42C. How much wattage, certainly well over 500W because even a quick analysis of the chart will show a 100W or greater jump in dissipation by allowing each additional 5C of delta-T with anything other than an very low speed fan. Which should make everyone rip off those "silent" fans and man up with some decent rad airflow and just turn the music up a notch. But to each his own, I am likely hearing impared from a Super Tramp concert where I stood next to the stage speakers - others may be more sensitive to noise. Which is why I always recommend med or high speed fans on a controller - dial in your temp to noise tradeoff to your desires.

http://www.skinneelabs.com/Radiators/XSPC/RX120/RX120-main.html

My point is that almost everyone I see post here goes overboard in suggesting size and amount of rads (which is not necessarily a bad thing if you have the cash and can figure out where/how to mount them.) because of some preconceived notion of what is acceptable for noise and temps or some rating by the rad seller. Very few of these companies actually make the rads they sell, certainly not the cores. As I mentioned before these things will dissipate more and more heat until they melt and they all work off of Delta-T, the difference between the air temp over the fins and the fluid temp and the bigger the difference (hotter loop - cooler air) the better they work. Which is why low speed super quiet fans are the worst thing you can do to a rad and people end up needing two huge rads to cool an almost insignificant heat load when compared to what the rad could dissipate. But quiet is good too. Read my advice to the OP again. It spells out the trade-offs and leaves the decision to him.

One may not be happy with the noise the required fan(s) makes. One may not be happy with the idle temps but the ability to cool a 1000W CPU with a dual rad within the operating temp envelope of say a C2D is certainly possible and if they keep adding cores and on-board memory controllers we might even be there in a few years.
 
Last edited:
Now for a really off the wall question. Do you think that the OP would be helped by using the Primochill Typhoon III Dual Bay Reservoir (I don't think it would be as good as having two independent loops). Could the OP put the 2x GPU's in one parallel loop and the CPU in the other parallel loop (maybe adding another rad to the GPU loop :confused:) or would the eventual mixing of fluids finally negate any advantages (even with a added rad) :confused:

Depending on the answer a Primochill Typhoon III would save you the money of having to purchase another pump, rez, tubes, etc. I also think it might be easier and space saving :confused:
 
You must log in or register to reply here.
Back
Top