Nvidia GTX 1080, 1070 'Founders Editions' Just Reference Cards

[JackNaylorPE]

Yeah, you nailed it. I'll be upgrading this year and I'll be looking for a reference cooler for this very reason.

You have measured this ?....Twin MSI 970s w/ 26% OC top card operating at 68C here.

Lot of misinformation, does any read actual test reports ?

https://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_980_Ti/35.html
Reference 980 Ti at max OC = 84C

Idle temperatures are fine; idle-fan-off would have certainly been possible. Under load, the card reaches its thermal limit of 84°C after a minute of gaming or so, which will cause it to lower Boost clocks, and, as such, performance.

https://www.techpowerup.com/reviews/MSI/GTX_980_Ti_Gaming/34.html
MSI 980 Ti at max OC = 75C

Idle temperatures are fine with even the idle-fan-off feature we love so much. In gaming, the card does not get anywhere close to the 84°C temperature limit at which NVIDIA's Boost will start reducing clocks to keep the card at a maximum of 84°C.

https://www.techpowerup.com/reviews/Gigabyte/GTX_980_Ti_G1_Gaming/34.html
Gabyte 980 Ti at max OC = 71C

Idle temperatures are fine, even with the idle-fan-off feature we love so much. In gaming, the card does not get anywhere close to the 84°C temperature limit at which NVIDIA's Boost will start reducing clocks to keep the card at a maximum of 84°C.

1. So much for the reference designs running cooler. Do the math ... the 84C of the reference design running at its throttling point is NOT cooler than 75C or 71C

2. The cards throttle at 85C ... yeah go out and buy those reference designs, that throttle when gaming.

3. Hybrid coolers are bringing nothing to the table. The aftermarket cards have no trouble with GPU temps, if a card is going to be OC temp limited, it's going to be VRM temp limited. Hybrids don't help here.

4. It must be said that nVidia has been upping the quality of their PCBs in recent generations. With the 9xx series, both Gigabyte and MSI used the reference Voltage controller that Nvidia used on their reference card. Both Gigabyte and MSI used the reference Hynix memory that Nvidia used on their reference card. So what we are seeing is that the non-reference cards are either the same or just slightly modified .... certainly to a lesser extend than in previous generations.

What remainss significantly different is how they cool the chips, sometimes how many power phases they supply, and of course the methods of cooling.

 

[xyriin]

Radiators shouldn't pull heated air into them. That is akin to placing the bomb inside the bomb proof bunker. The purpose of the idea is defeated. A custom water loop will normalize and any heated air introduced into the case should have no significant impact in raising internal case temperatures. Experiencing significant rise of internal case temperatures can include insufficient exhaust fans, not enough radiator space for the cooling requirement, too little flow (gpm) through the loop(s), insufficient air CFM and/or insufficient fan air pressure to push enough air through the radiator(s). I am sure there are more reasons/causes that escape me at the moment. The best case (no pun intended) scenario is to move the rads into a rad box, into a pedestal, or otherwise outside the case to intake cooler external air and exhaust this air back into a larger external space.
If running a custom water cooled loop, the GPUs will most likely be under water like the CPU (& vice versa); the idea it is necessary to add a significant number of extra fans as pure exhaust is not realistic.

Both non-reference cooling solutions and reference blower style coolers, pick up hot air from inside the case to cool a hot running graphics chip. This is rather counter-intuitive from the start. Both types of solutions exhaust a certain amount of that hot air inside the case as well. A non-blower type solution will obviously dump a larger volume of air back into case. This is why proper air flow through a case is so important. A case and/or fans many not allow/provide enough air to flow through the case necessary to provide effective cooling, not to mention failing to eliminating dead air spots and properly exhausting the "used" hot air.
I have two 18 year old Antec cases (model 1080 AMG - one stock - one modded with dual top exhaust fans) utilizing only 80mm fans and both keep air cooled Crossfire/SLI configured GPUs within 2° - 3° F and allow for "normal" OCing on air. The modded case is about 3° to 4° F cooler in general.
My main rig consists of a CaseLabs Merlin ST10 on top of dual pedestals. One pedestal mounts two, 480mm rads and the other pedestal mounts two, 360mm radiators. Of course there is no problem with radiators dumping heat into the case. Prior to expanding the system into the pedestals, the ST10 case housed three, 360mm radiators in push-pull, all used as air intake with two 120mm exhaust fans – the power supply and one other fan. This caused no adverse air temperatures inside the case.

Makes a lot of sense, because unless you're using some crappy old case or one with very little cooling support, there should be no reason to not be using a sufficient quantity of quality fans in exhaust configurations that exchange the heated air inside the case, to the outside of the case, several times per second under high RPM fan operation. Since the card will be cooling with the water loop rather than using internal case air, it should be a non-factor.

The only time using a radiator configured for intake should be an issue is if you don't have adequate case cooling for whatever reason, most especially in the exhaust locations. As to how this might affect CPU cooling, depends on the type of cooling used for that, the case model and how many OTHER intake fan locations you have in addition to those being taken up by your GPU card loop.

You're missing the scientific principle behind a radiator. Heat removal by a radiator is a function of delta T (or difference in temperature). There is never a single correct answer for any setup, as often the #1 factor in intake or exhaust is what type of cooler your GPU(s) are using.

In an intake radiator setup your cooling air is room temperature (we'll use 70F). Now let's use 70C (158F) for the CPU temp. Assuming you get perfect equilibrium (doesn't happen but we're trying to keep this simple) you're dumping 114F exhaust air into your case. All this hot air will of course raise the ambient temperature inside your case and an increase in ambient temperature inside the case drives up the baseline temp of all your components. The way you counteract this in the intake setup is to increase your airflow in an attempt to move the hot air inside the case out of the case.

In your bomb example the 'bomb' is the heat from the CPU. The intake example actually takes that bomb from the CPU and dumps it right back into the 'bunker'...your case. You then have some poor sap trying to run the bomb out of the bunker (case fans).

Now, take a properly configured exhaust setup. You've still got good airflow through the case and because you're not dumping the 'bomb' back into your case your ambient temperatures are going to be very very close to room temperature. Yes, they will be a few degrees above that 70F air you're running through the radiator in the intake example so your exhaust radiator is going to be a few degrees lower on the delta T. While this seems like a loss, you're forgetting that you're dumping the bomb directly outside the case and your ambient temperature is barely above room temperature.

Testing shows that an intake exhaust configuration leads to slightly lower CPU temps but it also leads to HIGHER GPU temps (again because you're dropping the CPU bomb in the bunker). However there is a massive variable I hinted at above, your GPU coolers. If you're using a reference style GPU cooler then the exhaust setup will be superior because the GPU isn't dumping exhaust heat into the case, it's dumping it outside the case leaving your ambient temps low. However, if you're using an aftermarket GPU cooler that dumps heat into the case then the intake setup is superior, at the expense of your GPU temps. Now if you also water cool your GPUs then yes, intake will once again pull ahead. However, you can't look at these things in a vacuum, configuration matters.

The other problem is that if you're utilizing multiple GPUs the reference design offers better cooling than the aftermarket coolers that dump the GPU heat directly into the case because of the clearance issue.

My setup is a Corsair 760T case, 2x 980 GTX in SLI and an i7-3770K overclocked to 4.4GHz (1.27V). All this is cooled with 2x 140mm intake fans (front), 1x 140mm exhaust fan (back), and a 280mm CPU radiator with 2x 140mm exhaust fans (top). All of my fans are Cougar 140mm HDB fans running at 1200 RPM and a whisper quiet 19dB (great static pressure so I don't have to run a push/pull config on the radiator). Under Prime I'm sitting at a comfortable 67C average across all 4 cores, peak core is 69C.

 

[TJ Hooker]

In an intake radiator setup your cooling air is room temperature (we'll use 70F). Now let's use 70C (158F) for the CPU temp. Assuming you get perfect equilibrium (doesn't happen but we're trying to keep this simple) you're dumping 114F exhaust air into your case.

Bit of nitpicking here, but you can't just average the two temperatures to get the exhaust temperature. That's not how thermodynamics works.

 

[xyriin]

In an intake radiator setup your cooling air is room temperature (we'll use 70F). Now let's use 70C (158F) for the CPU temp. Assuming you get perfect equilibrium (doesn't happen but we're trying to keep this simple) you're dumping 114F exhaust air into your case.

Bit of nitpicking here, but you can't just average the two temperatures to get the exhaust temperature. That's not how thermodynamics works.

You are 100% correct, but I was just trying to get the principal across in an easy to understand manner for people who haven't had any classes in thermodynamics/HTFF/etc. It's also why I said that kind of perfect equilibrium doesn't happen. 😉

 

[JackNaylorPE]

You're missing the scientific principle behind a radiator. Heat removal by a radiator is a function of delta T (or difference in temperature). There is never a single correct answer for any setup, as often the #1 factor in intake or exhaust is what type of cooler your GPU(s) are using.

There is only one scientific principle in play and that is the Law of Thermodynamics.

1. Yes, Delta T is the factor which will determine the effectiveness of the cooling

2. Yes, there is a single correct answer for every setup. The cooler the INTAKE air is, the more effective the cooling solution will be... no exceptions. You bought the water cooler for what ? To keep CPU temps down because, when overclocked to the level you want to run at, you are concerned about the temp of your CPU. Is there any other thing inside your case that you are concerned about reaching a hi temperature ? If there was, I can only assume you would water cool that too.

3. Logic dictates that interior case air temps will be warmer than exterior air temps. Therefore there can be no exceptions. Using cooler exterior air will always produce lower CPU temps than using hotter interior case air. As far as the impact on interior components, logically there is an impact. But the simple fact is when ,measured, it is small. At least that's what we have measured in every build we have done.

In an intake radiator setup your cooling air is room temperature (we'll use 70F). Now let's use 70C (158F) for the CPU temp. Assuming you get perfect equilibrium (doesn't happen but we're trying to keep this simple) you're dumping 114F exhaust air into your case.

No where near it.

http://www.iig-llc.com/blog/2015/02/too-hot-to-handle/

The threshold of pain for touching any surface inside your PC is 111 F. And yes if their air inside your PC is 114F, all surfaces within the PC will be 114F.

We are talking about a water cooling system and this is the "reality". These are real measurements taken just seconds ago on an overclocked 4770k (4.6 Ghz .... 1.387v) with core temps ranging from 69 to 74C and 6 temp sensors with CPU at 100% load under RoG Real Bench

Intake Air = 21.8C
Interior Case Air = 25.7
Radiator In Water Temp = 29.1
Radiator Out Water Temp = 28.6

So as we see, 25.1 C is not anywhere near 114F, its barely above room temperature on a warm day at 77C. Is there a concern about any component within the case being affected by 25.1C air ?

It comes down to this

Delta T Intake = 29.1 - 21.8 = 7.3C
Delta T Exhaust = 25.7 - 21.8 = 3.9C

The radiator / CPU cooling is 1.87 times as effective as intake.

I can't get a 980 series card to break 70-75C "on air" so with a throttling temperature of 84C, what is our concern ? According to techpowerup....

Gigabyte hits 66C OCd
Asus Matric hits 61C OCd
MSI Gaming hits 67C OCd

Add 10C for the top card (assuming you didn't buy a proper MoBo and nor install a side panel or back of HD cage fan) ... are we throttling yet ? Nope, so why do we care ?

Testing shows that an intake exhaust configuration leads to slightly lower CPU temps but it also leads to HIGHER GPU temps (again because you're dropping the CPU bomb in the bunker). However there is a massive variable I hinted at above, your GPU coolers. If you're using a reference style GPU cooler then the exhaust setup will be superior because the GPU isn't dumping exhaust heat into the case, it's dumping it outside the case leaving your ambient temps low.

No, while it may seem logical at 1st glance, this is not supported by testing. Ever review always show the blower style coolers with a hotter GPU. So the CPU tests show the CPU cooler with intake fans and the tests show conclusively that GPUs run hotter on blower style coolers and, .... lets not forget, blower style coolers are running at lower OCs. Let's not forget that the cards with blower style coolers ARE THROTTLING in the published tests. The open coolers are 17 to 23C below the throttling point.

Again, we uses 6 temp sensors, infrared thermometer, fog machine, 6 channel temperature display (which displays 0.1C accuracy of sensors), and HWiNFO64 as well as other utilities.

First off the GPUs are NOT dumping all that air outside your case. There is plenty of air leakage from the cooler shroud ..... then there's also thermal radiation.

1. When you design a radiator system, the radiators are designed to dissipate just 60% of the heat load of your CPU / GPUs (and whatever else is water cooled). That means that 40% is radiated from component enclosures, shrouds, PCBs, tubing, backplates and other surfaces inside your case. Those 2 x 250 watts GPUs and OCd 125 watt CPU ... w/ all water cooled and ignoring everything else, you have 250 watts heating air inside your case from thermal radiation and 375 being cooled by the radiator.... and that's not even counting PSU, Optical (10 watts), HDs (5-10 watts), RAM , MoBo (40 watts) , water pump (20 watts) whatever

2. Your blower cooler does not exhaust 100% of the air outside the case. The blower coolers have their place ... that being non-overclocked small ITX builds w/ inadequate case air flow where you have nothing else to exhaust that air. But in mid size case you not only have the case fan inlets and outlets you have the case grilles. Your design goal should be at worst, one 140mm, 1250 rpm fan for every 100 watts of heat generation.

3. One mistake many folks make is trying to balance intake an exhaust fans in the idea that "equilibrium" is a good thing ... it's not. Doing so, eliminate on of the cases best cooling features ... open grilles. Using radiator fans as intakes you are doing 1 of 3 things:

a) You have more fans blowing in than out resulting in slightly more air coming in than going out thru fans. This means you have slight positive pressure (good for keeping dust out) and you are pushing a small amount of air outside thru the case grilles.

b) You *think* you have the same air blowing in as out because you have an equal number of intake and exhaust fans. The reality is that you have a lot less air coming in thru fans than going out because you lose about 1/3 of the air flow on intake fans due to the resistance of an, even a freshly cleaned, air filter on the intake fans.

c) You have more fans blowing out than in which is the absolute worst possible conditions

If I am reading correctly, you have two intake fans in front, two exhaust fans on top and a rear exhaust fan . Considering the intake filters, that means that you have about 2.25 as much air blowing out, than blowing in. As a result, your case must pull in 1.25 times worth of air as intake thru case grilles. That means...

a) a lot of air is coming in thru unfiltered openings, covering components with dust and acting as a heat insulator.

b) where is that air coming in ? The most common place is the rear slots and case grilles. Remember those blower style coolers exhausting all that hot air outside the case ? You have created a large intake air deficiency and your air starved case has few options to suck the necessary air into the case. The most likely source is coming right back in thru the rear case slot grilles.

Your case also has a big wide grille above the exhaust fan. All that blower air and case air coming out thru the rear fan has to pass by that big wide grille above the fan.... and, due to the negative pressure inside the case, will come right back in.

So after going thru all that effort to get all that hot air out of the case with those blower style coolers, your decision to use the rad fans as exhaust fans results in all that blower style hot air exhaust being sucked right back inside the case instead of nice, fresh cool ambient air.

If you have sealed those grilles up, the effectiveness of all your fans is greatly reduced. The fans will operate well below spec due to the negative pressure inside the case. Air flow will be cut drastically and SP will be significantly reduced having to overcome that negative pressure condition because you case is intake air starved.

Invest in one of these:

http://www.amazon.com/Chauvet-Hurricane-901-Fog-Machine/dp/B008HZEW54/ref=sr_1_4?ie=UTF8

... and this will immediately become evident.

Lets switch it around.... with 4 fans blowing in and 1 out ... you are not sucking any exhaust air air in thru unfiltered case grilles and you have eliminated the dust problem. All the air exhausted by the rear fan and blower style coolers will exit the case and .... now ... stay outside the case. It also must be recognized that heat doesn't "build up" inside a case because air is being turned over quite fast. It equalizes very quickly due to the large air turnover.

Your case is 22.40" x 9.70" x 22.20" or 2,79 cu.ft. Assuming ya have components eating up 20% of that volume, we are left w/ 2.23 cuft of air. Your case fan specs, say the fans push 90 cfm of air. That is not true... nor is it true that they are 19 db at full speed. Those are reported on a test bed with no air resistance. With an inlet filter in front or radiator behind actual flow is a lot less and noise is a lot more. For the sake of argument lets say they push 50 cfm as the lower it is the more it hurts my position.

5 fans x 50 cfm = 250 cuft per minute

That means that the air inside the case is being turned over with fresh air 112 times per minute or 1.87 times EVERY second. In other words that 77F air is replaced with 71F air every 0.53 seconds. Again, fresh ambient air, not preheated freshly exhausted air.

The other problem is that if you're utilizing multiple GPUs the reference design offers better cooling than the aftermarket coolers that dump the GPU heat directly into the case because of the clearance issue.

There's only a "clearance issue", if you have chosen your components poorly. Yes, manufacturers sell inexpensive boards that are advertised as "SLI capable". The chipset is certainly SLI capable but that's not where the selection criteria should end. They also sell 4 wheel drive cars that should never be taken "off road".

When building 2 x SLI:

a) Pick a larger PSU
b) Choose a case that has fan mounts on the back of the HD cage or on the side panel to blow air between the cards
c) Choose a MoBo that has an empty slot (or 2) between the cards so that your top card is not intake air starved. Without this spacing 1) what air it does get is preheated by the bottom card and, more importantly, 2) you card is intake air starved because of the limited spacing between the cards .. which will be much worse w/ a blower style cooler.

The "clearance issue" by no means goes away with blower style cards because you have made the same mistake you did with the fans. Yes the blower style card blows more air outside the case but it can not do that w/o a sources of intake air. The teeny space between the cards limits air flow and what greatly reduced amount it does get is substantially preheated by the thermal radiation and escaped air coming off the card below. Your cooler cools the GPU but all the heat coming off your PCB is going where ? Up into the card above. The "tight" shroud design provides much more air intake restriction than the "open design" so any advantage it might gain by pushing a higher % of air out is erased by its inability to get fresh cool air in.

 

[USAFRet]

There is also one other method of determining optimal rad fan config ...intake or exhaust...for you particular rad in your particular case, components, and cable management situation.

Try it both ways.

However the system is now, record the temp. Idle and under load after 30-60 minutes
Record this for the CPU, GPU, and everything else.

Now...swap the fans around, and test and record again.

Whichever is better, for you, in your case....do it that way.

In my AIR 540, with a Cryorig A80, blowing out...it is sufficiently cool enough that I do not feel motivated enough to take it apart and flip the fans over.

Currently, at 24C ambient room temp, I get ~34C idle, and 49C under 100% load (30 minutes with Prime95)


All of this theoretical positive/negative/intake/exhaust is completely negated by a hamfisted cable management job, or drives blocking a fan, or that fancy lighting, or, OH WAIT, its sitting on carpet, or the PSU is installed upside down, etc, etc, etc.....
What works in your box?

 

[xyriin]

2. Yes, there is a single correct answer for every setup. The cooler the INTAKE air is, the more effective the cooling solution will be... no exceptions. You bought the water cooler for what ? To keep CPU temps down because, when overclocked to the level you want to run at, you are concerned about the temp of your CPU. Is there any other thing inside your case that you are concerned about reaching a hi temperature ? If there was, I can only assume you would water cool that too.

If you're only talking about the CPU then maybe, then again there is a major factor you're ignoring. I'll take an intake temp of 22C, you get 20C. By your logic you've already won the cooling war. However, what I didn't tell you is that even though we're cooling the same CPU generating the same heat, your CPU is placed in an enclosure 5C warmer than mine meaning my OPERATING temperature is cooler so you lose.

This is why dumping all the heat back into the case DOES affect your operating temperature. If you immediately dump all the heat from the CPU radiator immediately out of the case it's not added back to your system. If you dump it into the case law of conservation of energy says you didn't cool the system at ALL. To overcome the CoE you increase air flow to move that heat out of the system.

3. Logic dictates that interior case air temps will be warmer than exterior air temps. Therefore there can be no exceptions. Using cooler exterior air will always produce lower CPU temps than using hotter interior case air. As far as the impact on interior components, logically there is an impact. But the simple fact is when ,measured, it is small. At least that's what we have measured in every build we have done.

Yes but how much that difference is can certainly be a a factor. Obviously if your ambient case temperature is way above room temperature (poor airflow and/or case discharge GPUs) then you're going to get poor returns on an exhaust radiator.

Intake Air = 21.8C
Interior Case Air = 25.7
Radiator In Water Temp = 29.1
Radiator Out Water Temp = 28.6

It comes down to this

Delta T Intake = 29.1 - 21.8 = 7.3C
Delta T Exhaust = 25.7 - 21.8 = 3.9C

The radiator / CPU cooling is 1.87 times as effective as intake.

Yes your intake Delta T is 7.3C, what you calculated wrong is the Delta T exhaust. In your intake math you're properly looking at the water temp versus air temp, the two mediums transferring heat in the radiator. If you moved your radiator to exhaust the correct math would be 29.1 - 25.7 which is 3.4C. Right now you're probably cheering because the gap widened, except for the fact your difference in air temps only seems like it's 3.9C. The catch is that your intake method is skewing your interior case air temperature because you're adding heat to the interior case with your intake radiator.

As for the difference in GPU coolers I won't disagree with you that the interior case exhaust models can be clocked higher or run cooler. The problem is that they once again heat up the interior of the case. And while you can argue it's all about board arrangement for SLI what about more than two cards? Eventually you're running out of PCIe slots you can skip. Yes there is radiated heat and some leakage but it's directly removing more heat from the case than any other form of GPU cooler unless you're using liquid on an exhaust radiator again.

As an additional side note related to pressure, I'm sure you're aware that radiators themselves affect flow and on top of that fan speeds can be adjusted as well. All that aside it's very simple to perform a physical test to see if you're positive or negative in the case and from where any leaks are happening...fog machine is nice but overkill unless you're doing engineering design.

The final fact you're completely ignoring which I've iterated on a couple times now is the noise and airflow factor. I don't want a million fans and a wind tunnel. I mean, that's the other bonus to liquid cooling after all. Sure it's great for overclock in those few times you actually use it, but 99% of the time it simply reduces noise level for you. I don't want to throw away that advantage by doubling the noise in my system. I used to use an intake exhaust setup exclusively because at first glance it made more sense. However, after using some math it didn't make that much sense in light of ambient temperature and noise. Even look at your example...your intake radiator is only dropping the temperature by 0.5C across the radiator! Even if you use your 1.87 effectiveness factor (which is skewed) you're only talking a fraction of a SINGLE DEGREE difference in the radiator drop! You're simply paying a higher overhead in ambient temperature and noise.

All said and done we obviously have a philosophical difference. You're not going to sway me though because I used to run an intake radiator even back when I had a single 1x 120mm radiator. Two builds ago I decided to do a test and for my configuration I lowered my CPU operating temperature with the exhaust setup. Yes my CPU cooling effectiveness dropped by just over a degree but the drop in overall case temperature dropped the baseline by almost 3 degrees which resulted in a 1.5-2C drop in the CPU's realized operating temperature at peak load.

 

[doomtomb]

look. somebody worked really hard to make this new cooler look like a corvette. give nvidia a pay day, they need it!

 

[Darkbreeze]

@xyriin, water cooled systems are ALWAYS louder under serious loads. That's just a fact. Case fans connected to sys or cha fan headers rarely if EVER ramp up to the speeds that the fans on a radiator do, when connected to the CPU or CPU OPT headers. So any liquid cooled system, that is reacting to an entirely different sensory hardware table based on CPU thermals and controlled via the CPU headers, is going to run those fans at high speed a lot more often, and at a much higher sustained RPM, than any two case fans running off a controller or the chassis fan headers, meaning it's going to be much louder. The fact that those fans are going to be sucking or blowing air through the additional restrictions created by the radiator fins also means the airflow, even at an approximately equal CFM and static pressure rating, will also be louder than a standard configuration of a case fan moving air through only the restrictions of the case vent itself. So much for your noise factor.

And anybody not willing to use a sufficient number of case fans, in order to move a high quantity of air at a lower RPM, resulting in a decreased level of noise, isn't too worried about it in the first place. You don't have to have a "wind tunnel" in order to keep the air inside the case substantially close to ambient. Also, a CPU fan buried inside the case is discernibly quieter than a radiator fan with less than an 1/8" between it and the outside of the case where noises are commonly more audible.

I think your assumptions are all off base, and I believe Jack has already shown and proved to you why that is. But I seriously suspect you'll not see it that way, so I guess it's pointless to argue these points further. Horses, water and all that.

 

[xyriin]

My radiator fans are low RPM and a mere 19dB, marginally above background noise in a quiet room. You simply use larger fans (140mm instead of 120mm/80mm so that you can run at a lower RPM making less noise while moving the same amount of air). Also why would I mount my fans on the case side of the radiator when that just increases the noise level. As a result I've got more than an inch between the fan and the case (radiator thickness). Utilizing quiet fans to begin with also helps with the noisy airflow issues...it's just a simple engineering problem with fan blade design.

I've already outlined my system specs and cooling configuration. I've got a whisper quiet system, keep temps in the mid 60s under a full load on a 4.4GHz OC, do it with an exhaust radiator, and to top it off my case ambient temp sits about 1-2C above room temperature.

I don't think you've really ever tried to optimize a fast and quiet overclocked system before. My solution isn't just theory, it was trial and error finding the best solution. Granted I don't buy the cheapest fans available, but I'm not spending a fortune either.

Now I could increase the noise level of my system to buy another 5C+ on CPU temp but why? I'm not going to be pushing it above 4.4GHz.

As for Jack he made some good points, but also a lot of assumptions specific to his configuration. USAFRet actually had the correct answer...there are multiple answers and it depends on the configuration.

 

[Syed Khalid Hasan]

win7 motherboardP4P800-VM sound driver problem

 

[USAFRet]

win7 motherboardP4P800-VM sound driver problem

Please create a new thread for your issue.
Thanks.

 

[refillable]

I don't know but if they're just really reference cards then it's going to be very silly. But do normal reference cards fare better in terms of Overclocking than Aftermarket cards? I don't think so...

 

[lieutenantfrost]

How long do you think before aftermarket competitor cards will be available?

 

[TechyInAZ]

How long do you think before aftermarket competitor cards will be available?

Probably a few weeks after the founder cards come out. That's typically how it goes. But with the price change in the reference design, it could be different.

 

[littleleo]

11 days to go will there actually be GTX 1080 stock to ship or not is the question.

 

[rhysiam]

How long do you think before aftermarket competitor cards will be available?

Probably a few weeks after the founder cards come out. That's typically how it goes. But with the price change in the reference design, it could be different.

I'm guessing it'll be a longer delay this time around. The only reason I can see for the existence of this over-priced "founders edition" is that it enables Nvidia to grab higher margins when demand is high after release. I'm expecting that we won't see partner cards until Nvidia has milked the early-adopters dry. It'll be interesting to see though.

 

[alexandrosgr]

It seems to me that nVidia just pulled one on the whole world with this trick.

Why would any partner price their board at 600 when they can charge 700 for it? I highly doubt we will see the 379 and 599 price points until AMD comes up with the 1070 and 1080 counterparts.

 

[TJ Hooker]

It seems to me that nVidia just pulled one on the whole world with this trick.

Why would any partner price their board at 600 when they can charge 700 for it? I highly doubt we will see the 379 and 599 price points until AMD comes up with the 1070 and 1080 counterparts.

Partners have always been able to charge whatever they wanted. But competition forces them to price them reasonably. I can't predict what prices are going to look like, but at this point I think it's premature to start worrying about price gouging.

 

[TechyInAZ]

How long do you think before aftermarket competitor cards will be available?

Probably a few weeks after the founder cards come out. That's typically how it goes. But with the price change in the reference design, it could be different.

I'm guessing it'll be a longer delay this time around. The only reason I can see for the existence of this over-priced "founders edition" is that it enables Nvidia to grab higher margins when demand is high after release. I'm expecting that we won't see partner cards until Nvidia has milked the early-adopters dry. It'll be interesting to see though.

Good news is that the founders edition is still a high quality card. Just as capable as supposed future medium range GTX 1080s.

 

[Memhorder]

Radiator as Exhaust no exception. My CPU stays at 47C with OCCT. One thing you guys never touched DUST. Intake Rad will get clogged. I can still see through my rad after 10 months however my Filters are clogged after 3. Factor that in the thermodynamics.

Linus done a test there on Ref vs Non and it showed that the reference card does run hotter yet all internal components were an average of 5 degrees cooler. This was single 970's. SLI would bump it up some

 

[rhysiam]

Good news is that the founders edition is still a high quality card. Just as capable as supposed future medium range GTX 1080s.

It seems okay, but not great. Especially not for overclocking. From the TH review:

Nvidia's direct heat exhaust cooler does do its job, but the GeForce GTX 1080 Founders Edition does face clear and restrictive boundaries that make overclocking completely pointless for sustained or challenging loads. Admittedly, the load’s a lot lower if you dial back to 1920x1080. But who buys a $700 graphics card for Full HD?

With an OC, unless they ran the fan at 100%, the clock speed in a gaming run dropped back to stock within about 4 minutes to keep thermals in check.

So it's an adequate cooler for stock, but pretty useless for OCing, unless you can bear a fan at 100% while gaming. I wouldn't consider that high quality personally.

 

[littleleo]

If Nvidia is selling the reference cards for $699 and your a board partner and you running a GPU clocked at a higher speed with a more advanced cooling solution why would you sell it for $100 less? Why not charge more then the $699?

 

[rhysiam]

If Nvidia is selling the reference cards for $699 and your a board partner and you running a GPU clocked at a higher speed with a more advanced cooling solution why would you sell it for $100 less? Why not charge more then the $699?

My guess would be that Nvidia will only allow the release of partner cards once the vast majority of people willing to spend $700 on a GTX 1080 have already parted with their money. If partners want to sell in volume, they might not have a choice but to sell closer to the MSRP.

 

[littleleo]

If Nvidia is selling the reference cards for $699 and your a board partner and you running a GPU clocked at a higher speed with a more advanced cooling solution why would you sell it for $100 less? Why not charge more then the $699?

My guess would be that Nvidia will only allow the release of partner cards once the vast majority of people willing to spend $700 on a GTX 1080 have already parted with their money. If partners want to sell in volume, they might not have a choice but to sell closer to the MSRP.

I thought I read where Nvidia will let partners sell starting on the 27th when they release their cards, but it may take longer for the Board partners to perfect their models.