uglyduckling81 :
The Nvidia reference cards do throttle though with those blower style coolers.
The aftermarket ones don't but then they do drop their hot air into the case rather than taking it away like the water cooled ones do.
I agree there isn't much point to it but it does give the advantages of both reference blower designs and the superior cooling of aftermarket solutions. It's a 2 in 1 deal.
How can you argue that the blower design is an advantage, when
a) It runs hotter than a normally cooled AIB card ?
b) it can not come close to matching the performance of an AIB card ?
How can you claim this is a superior cooling solution when
a) it provides no advantage and
b) if installed
as per Corsair's published recommendations ..
.radiator fans are installed as intakes which means you two arguments are opposing one another. Why is Corsair publishing installation instructions saying to install rad fans as intakes ?
I repeat it so often that I didn't bother this time .... but isn't it obvious ? 1) An AIB is cheaper than an FE so who would buy it and 2) it has never been advisable to buy a reference card. There is a reason why blower style cards never test well, main one being blower style coolers don't cool well. No one has ever explained why blowing air into the case is an issue other than "heat is bad". What component are we worried about here ? Why do non-blower cards always have lower temps and better performance ?
Do the math ... a properly designed case does a complete air turnover on the average of
twice every second....what are we worried about here when all the heat is outside the case in half a second ? Do the math with blower style coolers and CLCs on CPU "exhausting air out"
contrary to CLC manufacturer's recommendations. What results is more air blowing out than blowing in which means that GFX card blower exhaust as well as the hot PSU exhaust is being sucked right back in thru
the rear grille.
Tradesman1 :
+1 Unless of course you set it up to blow air from the outside in, which will increase the heat in the case
On the other comment, the purpose of any cooling is to keep temps down....period...the lower the better, which can open the threshold for higher OCs and/or prolong the life of the component, so no, it's not pointless.
Better said, I think ....Unless, of course,
you install exactly per Corsair's published installation instructions:
For the best cooling performance, we recommend mounting the fans as an air-intake to your PC case.
As for the purpose of cooling, I think clarification is needed ?
The purpose of cooling is not to keep components within accepted operating ranges not to reduce cooling just for the sake of cooling "Bragging rights". Same for CPUs.... was a day that you reached the temperature limit long before you reached the voltage limit. Now, I seem to hit the voltage limit more often then we hit the temperature limits. If I am seeing peaks of 1.5volts and 72C, I see no reason to try and improve cooling... I'm not going to raise that voltage so a better cooler isn't going to do anything for me.
When I say pointless, there is no point in going to every increasing expense to reduce temperatures if it does not provide a return on the investment. To take a specific GFX card which his max stable OC at 74C is not going to get any higher clock at 50C
Cost must always be part of the equation, otherwise there's no room to consider anything but the best
1. The 1080 will start to throttle after breaking 82C .... any decent AIB card will be in the low 70s when at it's
max OC. So, no ... there is no potential for bigger OCs because no AIB card with a decent air cooler is getting anywhere near it's throttling point.
The days where liquid cooling brought anything to the table performance wise are gone, at least on the nVidia side. It's simply a matter of do I pay a small price premium for a decent air cooler or do I pay twice that premium or more for something that will provide zero return on that extra investment ?
2. There is no evidence that has **ever** shown GFX cards to be at any risk to useful life from running below their throttling point. If you want to kill a decent AIB GFX card, you need to overvolt and nVidia has made it virtually impossible.
On the 7xx series and even 9xx series, it was kind of silly as despite the 50C GPU temps, VRM temps were getting well into the high 80s and affecting stability. The 10xx series is free of this concern. However if we are going to argue that the GPU benefits in some way from the hydrid cooling then why has no one ever explained why the same liquid cooling logic doesn't extend to the VRM or memory ? VRM Failures are historically, the most common cause of GFX card failure.
We have a build here with two GTX 560 Tis in SLI ... the cards start throttling after 84C and they are just below that with a 28% OC. In a few months it will be 6 years old. In 23 years, we have never had a GFX card failure ... never heard of a failure related to an nVidia GFX card electromigration / heat degradation at any "safe OC" where throttling does not occur. Did read of AMD mining cards go down on occasion but those old cards are not under discussion here.
A hybrid is more at risk from galvanic corrosion than an air cooled 10xx AIB card is from thermal deterioration.
https://martinsliquidlab.wordpress.com/2012/01/24/corrosion-explored/
There certainly are reasons to water cool a GFX card ... aesthetics are better, SLI problems disappear, noise reduction is significant (at least on custom loops) but the simple fact is nVidia's cards have gotten so efficient, that water cooling brings nothing but "bragging rights" to the table.
My personal box has SLI'd 780s with Furmark temps of 39C if i don't limit fan speeds.... limiting max rpm to inaudible levels, results in 44C. The 5C sound gain, has a return on investment ... box is completely dead silent. The lower 5C, does not have one. On air, the card is just over 69C at same OC .. well below the 95C danger point and well below the temperature at which performance is in any way affected.