Why dual loops? Water cooling.

[Enderman]

1 minute ago, Jumper118 said:

when benching you dont let the water temperature get to a static temperature unless that temp is 0c. 

??

What the heck are you talking about?

Do you not understand what normalize means?

Obviously the water temperature is never static...

[Jumper118]

Just now, Enderman said:

??

What the heck are you talking about?

Do you not understand what normalize means?

Obviously the water temperature is never static...

no normalise is spelt with a s

[Enderman]

4 minutes ago, Jumper118 said:

no normalise is spelt with a s

http://english.stackexchange.com/questions/107902/normalise-or-normalize-british-english

Normalize is correct, sorry if you were taught wrong.

[Jumper118]

5 minutes ago, Enderman said:

http://english.stackexchange.com/questions/107902/normalise-or-normalize-british-english

Normalize is correct, sorry if you were taught wrong.

now your are just being a gammar nazi because you are wrong and cant win the argument.

[Enderman]

6 minutes ago, Jumper118 said:

now your are just being a gammar nazi because you are wrong and cant win the argument.

Really? Pretty ironic, since this guy who started trying to be right about grammar (and was actually wrong) is YOU!

 

15 minutes ago, Jumper118 said:

no normalise is spelt with a s

^^^

 

Oh, and that was your very smart response which totally had everything to do with how two loops do not perform better than one.

I would suggest you do a bit more research, both about your spelling, AND about how watercooling works, because you seem to have trouble understanding that how many loops, or where the radiators are in a loop, does not make a difference when all the water in a loop is at the same temperature +- a degree.

[Jumper118]

1 minute ago, Enderman said:

same temperature +- a degree.

and there it is  

[Enderman]

7 minutes ago, Jumper118 said:

and there it is  

There is what?

1 degree is not going to limit performance.

In fact, if you use a faster pump like a D5, that 1 degree will turn into 1/10th of a degree.

 

The way heat transfers from a component to a fluid is by the difference in temperature.

If the water is at 35C and the component is at 80C, having the water at 36C will not make a significant different to the delta T which is 44 degrees.

 

If someone cares that their CPU is getting 195W of cooling instead of 200W of cooling then they would probably invest in something better than watercooling, which is a pretty bad solution to getting top performance.

Which again, proves that people do dual loops for looks and not performance (except for the not-really-smart person that you posted a picture of above)

 

 

Also, good job at trying to avoid embarrassing yourself when you spelt "normalize" with an s  that was really funny

[Eddie Current]

11 hours ago, jasonvp said:

Simple: demonstrably better cooling performance...

@Jasonvp, 've done similar tests and have found your results to spot on. Something else some people over look is that with more volume and longer fluid line the longer it takes for components to heat up and might even allow for lower temps.

[SilentKnight777]

On 12/13/2016 at 7:04 PM, Eddie Current said:

@Jasonvp, 've done similar tests and have found your results to spot on. Something else some people over look is that with more volume and longer fluid line the longer it takes for components to heat up and might even allow for lower temps.

Old discussion. But just to throw in: 1) it doesn't make any statistical difference, and is primarily cosmetic.  2) if it DID make a difference, that difference would be variable, and often favor a single loop (depending on ambient temps).

 

why? Simple--radiator availability.

 

lets say you have a dual loop. Your CPU (and all the fluid in the loop) have access to only 1 radiator. Your GPU (and all the fluid in that loop) have access to only 1 (or maybe 2) radiators. 

 

so what happends when you run task 1: very gpu dependent? Your GPU temps rise, the GPU radiator loop can't keep up. Meanwhile your CPU stays nice and cool--the CPU radiator is taking a vacation. 

 

Or, inverse that for a CPU intensive task. The CPU loop gets hot, as it's radiator is overwhelmed, while the GPU loop takes a break and remains cool.

 

you have over compartmentalization of your loop, and any time one loop is cooler than another, you have misallocation of cooling potential. Assuming 2 equally sized radiators, 1 on CPU loop, the other on GPU loop. If the CPU loop is 38 degrees, and the GPU loop is 42 degrees, if you ran a single loop you would end up with an average temp closer to 40 degrees without the large variance. (Actually that depends a lot on GPU work, maybe closer to 40.5-41, but cooler for sure)

 

bottom line: single loop runs ALL your water, through EVERY radiator, giving you the MAXIMUM cooling for the ENTIRE loop. Compartmentalizations of your loop only lead to inefficiencies of cooling. You may get (and likely WILL GET) one of your loops cooler (likely CPU), but at the expense of your other loop warmer (likely GPU). Given your GPU temps are usually the rate limiting step--this is a mistake performance wise. 

 

I'm debating a dual or even triple loop for aesthetics, but I'm under no delusion I will be improving performance, likely the opposite...