GPU Temperature

[Aberforth]

I'm starting to get confused about G80 temps, Which temp is more important when ocing? Is it...

GPU
GPU Diode
GPU Ambient or
GPU Memory ?

And what are the danger zone temps for 8800 GTS and GTX?

Thank you

 

[BFG10K]

Temperatures aren't really your concern when overclocking; what?s important is whether you have stability issues or rendering artifacts.

 

[Aberforth]

i don't have any issues though, i'm just worried whether high temperatures are really screwing up my cards slowly. GPU diode temperature shows 86°C.

 

[JPB]

Since you say you have no artifacting or stability issues :

GPU= Mostly important while Core overclocking
GPU Diode = Not sure about this one
GPU Ambient or = Nice to have plenty of case air flow while overclocking. If you do, disregard this one.
GPU Memory ? = Keep an eye on while memory overclocking

When I had a GTS 640, Idle temps were 55, and load temps were 70 at highest.

What are your temps ?

 

[Aberforth]

well..I'll tell you about GPU Diode becasue thats the temperature shown in Nvidia Control Panel.

Geforce 8800 GTS

Idle : 69°C
Load: 86°C

 

[AmberClad]

Are those the stock temperatures? Or after the OC? If that's stock, that's rather high. My stock temps are around the same that JPB had (55C idle).

 

[n7]

Thanx to nVidia's advanced cooling techniques, my 8800 GTX is presently idling at a nice cool 80C

 

[Aberforth]

Originally posted by: AmberClad
Are those the stock temperatures? Or after the OC? If that's stock, that's rather high. My stock temps are around the same that JPB had (55C idle).

no they aren't stock...but look at this: http://img442.imageshack.us/img442/4376/vpskzawle7.png

 

[JAG87]

your temps are fine. gpu diode is the core temperature, it's like the core temperature in a cpu. The gpu is the temperature of the IHS. Nvidia monitor should be reporting the temperature of the IHS, not the GPU diode. Most software reports the IHS temperature not the diode temperature. The diode can get very hot, dont worry about it. Just make sure your IHS temperature is around 60 idle and 80 load. A few degrees hotter is acceptable too, like 65 idle and 85 load.

n7, you seriously need to look into an HR-03. I dont know what your setup is like, but 80c idle is excessive, you are torturing that card.

 

[Aberforth]

Originally posted by: JAG87
your temps are fine. gpu diode is the core temperature, it's like the core temperature in a cpu. The gpu is the temperature of the IHS. Nvidia monitor should be reporting the temperature of the IHS, not the GPU diode. Most software reports the IHS temperature not the diode temperature. The diode can get very hot, dont worry about it. Just make sure your IHS temperature is around 60 idle and 80 load. A few degrees hotter is acceptable too, like 65 idle and 85 load.

n7, you seriously need to look into an HR-03. I dont know what your setup is like, but 80c idle is excessive, you are torturing that card.

thanks, btw i'm getting HR-03 as well.

 

[bryanW1995]

I think he is torturing his ambient more than the card. I thought my 1950xt idled hot at 61c until I figured out the magic of 2d/3d clocks. It now idles at 47c . Load can get a little high, I think it gets to 86c or so running 3dmark06, but that doesn't seem excessive compared to what I've seen from others.

 

[MetaDFF]

By default, with automatic fan control, the 8800GTX idles pretty hot. I have a stock 8800GTX and it's idling at 73C reported by NVMonitor. Under heavy load such as artifact scanning with ATITool it can easily reach 88C with the fan at maximum speed. The airflow in my case is pretty decent with two 120 mm fans. Even with the case open, the temps stay within +- 1C.

 

[Aberforth]

Originally posted by: MetaDFF
By default, with automatic fan control, the 8800GTX idles pretty hot. I have a stock 8800GTX and it's idling at 73C reported by NVMonitor. Under heavy load such as artifact scanning with ATITool it can easily reach 88C with the fan at maximum speed. The airflow in my case is pretty decent with two 120 mm fans. Even with the case open, the temps stay within +- 1C.

I think I understand how that happens, nvmonitor reports incorrect temperatures if your mainboard is based on Intel Chipset, I have Intel 975xbx2 and nvmonitor reports 73°C, try checking the temps in everest

 

[LOUISSSSS]

Originally posted by: Aberforth

Originally posted by: AmberClad
Are those the stock temperatures? Or after the OC? If that's stock, that's rather high. My stock temps are around the same that JPB had (55C idle).

no they aren't stock...but look at this: http://img442.imageshack.us/img442/4376/vpskzawle7.png

hey what program is that called? it shoes ur gpu temps in-game?

 

[MetaDFF]

Originally posted by: Aberforth
I think I understand how that happens, nvmonitor reports incorrect temperatures if your mainboard is based on Intel Chipset, I have Intel 975xbx2 and nvmonitor reports 73°C, try checking the temps in everest

Just installed Everest and I don't see my GPU temperature listed anywhere in the "Sensor" section. I do see my CPU and HDD temps though.

 

[JPB]

I also just now installed Everest Home Edition. MY GPU temps are shown at 40c. Thats with a BFG 7900 GTX OC though.

 

[Aberforth]

Originally posted by: MetaDFF

Originally posted by: Aberforth
I think I understand how that happens, nvmonitor reports incorrect temperatures if your mainboard is based on Intel Chipset, I have Intel 975xbx2 and nvmonitor reports 73°C, try checking the temps in everest

Just installed Everest and I don't see my GPU temperature listed anywhere in the "Sensor" section. I do see my CPU and HDD temps though.

take a look: http://img399.imageshack.us/img399/637/hfxdhbkd6.png

 

[Yanagi]

I have an intel P35 chipset and nvidia utility shows my GPU temps as 55 idle, havent bothered to measure load temps.

Card is overclocked to 648/2100

 

[Sylvanas]

A while back there was a topic on AT which the general consensus was that the life expectancy of a CPU is 10years at stock frequency/voltage and changing frequency/voltage would knock off about 5 years (depending on situations). I would assume that if the Tjunction of the G80 is around 100c then wouldn't that mean Nvidia support that same kind of life expectancy of 10years @ their maximum operational temperature of 100c? I guess what I am trying to say is, I wouldn't worry too much about temps until something starts going wrong

 

[n7]

Originally posted by: JAG87

n7, you seriously need to look into an HR-03. I dont know what your setup is like, but 80c idle is excessive, you are torturing that card.

How am i torturing my card?

I am running case side open, stock cooling, default driver settings, default speeds.

If that's too hot, that's not my problem, that's a nV-being-retarded problem, as it's their cooling solution, & their driver settings which control fan speed.

So if their moronic settings & wimpy fan can't control temps to a decent level, they get the RMA when the card dies, as i've done nothing to deserve a malfunctioning card.

There's no reason why i should have to manually force fan speeds, as the fan speeds automatically kick in when temps increase slightly.

IOW, nV wants the card to idle @ 80C apparently.

 

[chizow]

Originally posted by: n7

Originally posted by: JAG87

n7, you seriously need to look into an HR-03. I dont know what your setup is like, but 80c idle is excessive, you are torturing that card.

How am i torturing my card?

I am running case side open, stock cooling, default driver settings, default speeds.

If that's too hot, that's not my problem, that's a nV-being-retarded problem, as it's their cooling solution, & their driver settings which control fan speed.

So if their moronic settings & wimpy fan can't control temps to a decent level, they get the RMA when the card dies, as i've done nothing to deserve a malfunctioning card.

There's no reason why i should have to manually force fan speeds, as the fan speeds automatically kick in when temps increase slightly.

IOW, nV wants the card to idle @ 80C apparently.

I find it somewhat ironic you take this attitude, yet you're the first to point the finger at NV when you run into problems.

As many have said, that's really hot for a G80. What's your ambient room temp? Simply taking off the side panel isn't going to cut it because it doesn't do anything to remove the hot air from around the card. If you want top-of-the-line performance with a part you know is going to run hot (by necessity) you should certainly make some attempt to keep it cool.

 

[n7]

I'm very quick to point fingers yes, because simply put, there's no reason i should have to worry about my card's temps being an issue.

If i was overvolting & OCing like mad, sure, but i'm not.

If they can't design a cooler capable of cooling the card adequately at defaults, that's not my problem.

BTW, if i set the fan to 100% manually, i idle around 65C.
Except @ 100%, it's too loud & annoying.

The thing is, the default driver settings basically attempt to keep the card idling around 80C.
As soon as it goes over that, then the fan speeds up a lot, so oddly enough, in games, it doesn't get much hotter than what it idles at.

Now i know most of you think this is harming the card, but if they write the drivers to not really kick the fan speeds up till i go over 80C, they must know what they're doing :roll:
(Or they're just stupid.)

So i'll leave that up to you guys to decide, but regardless, it's an nV issue, not an n7 issue.

 

[chizow]

Originally posted by: n7
I'm very quick to point fingers yes, because simply put, there's no reason i should have to worry about my card's temps being an issue.

If i was overvolting & OCing like mad, sure, but i'm not.

If they can't design a cooler capable of cooling the card adequately at defaults, that's not my problem.

BTW, if i set the fan to 100% manually, i idle around 65C.
Except @ 100%, it's too loud & annoying.

The thing is, the default driver settings basically attempt to keep the card idling around 80C.
As soon as it goes over that, then the fan speeds up a lot, so oddly enough, in games, it doesn't get much hotter than what it idles at.

Now i know most of you think this is harming the card, but if they write the drivers to not really kick the fan speeds up till i go over 80C, they must know what they're doing :roll:
(Or they're just stupid.)

So i'll leave that up to you guys to decide, but regardless, it's an nV issue, not an n7 issue.

Again, I'll have to disagree here. That's like buying ice cream at the grocer, which is nice and frozen when you buy it, then tossing it in your broken freezer and complaining that it melted. There's nothing wrong with NV's cooling solution, in fact many will say its the best stock cooler in terms of both performance and noise they've ever owned. Even ATI's new cooling solutions mimic it.

Reason I bring up the ice cream analogy is that there are default conditions and expectations for parts to run nominally. Even OEM systems you buy will have sufficient cooling to keep all parts running under normal conditions. There's a reason new cases continually improve cooling with each new iteration, because what was acceptable 3-4 years ago simply doesn't get it done anymore.

But hey, if you want to run your card at 80C and then scratch your head as to why you're having problems, yet do nothing about it (other than point the finger), that's your decision.

 

[n7]

LOL.

Ice cream?

Yes, i have an apartment with high ambient temps.

But as i'm telling you about the temps i'm running, it's on a frickin' open sided case.
Even if i had the best case in the world made of fans only, it's still only going to be blowing ambient temp air over things, which isn't going to cause my card to run cooler.
(And FWIW, temps are basically the same case side closed.)

The part you don't seem to get is that is incredibly simple.

It is not a user's responsibility to mess with fan speeds, add aftermarket coolers, etc. unless so wished.
When you buy a CPU, the stock fan is designed to run things just fine w/o issues.
Of course, if you add voltage & overclock, then it's ideal to purchase a better cooling solution, as you've now went beyond what the original cooler was intended for.

But in the case of my 8800 GTX, i am doing neither, which means there's no reason why i cannot assume the card runs quite fine using their default cooling & settings.

If they design a card with drivers that don't kick fan speeds up till it's 80C, that's a design decision.

You're trying to tell me that's my fault, but it's not.

They design the driver to only bump fan speed to combat temps when it crosses that temp threshold, which obviously means that's well within card's specs.
(Or that they're retarded, which i already stated, you decide).

If that wasn't "normal", they would have designed the drivers differently.

Get it?

It's not that hard to understand really.

It is not a simple user's responsibility to modify cards to run safely at default conditions.

Which is why i say it's obviously a safe temp to be idling at.

And if not, they've overlooked something, & potentially they end up with an RMA (assuming that's too hot, which it obviously isn't).

 

[chizow]

Originally posted by: n7
LOL.

Ice cream?

Yes, i have an apartment with high ambient temps.

But as i'm telling you about the temps i'm running, it's on a frickin' open sided case.
Even if i had the best case in the world made of fans only, it's still only going to be blowing ambient temp air over things, which isn't going to cause my card to run cooler.
(And FWIW, temps are basically the same case side closed.)

The part you don't seem to get is that is incredibly simple.

It is not a user's responsibility to mess with fan speeds, add aftermarket coolers, etc. unless so wished.
When you buy a CPU, the stock fan is designed to run things just fine w/o issues.
Of course, if you add voltage & overclock, then it's ideal to purchase a better cooling solution, as you've now went beyond what the original cooler was intended for.

But in the case of my 8800 GTX, i am doing neither, which means there's no reason why i cannot assume the card runs quite fine using their default cooling & settings.

If they design a card with drivers that don't kick fan speeds up till it's 80C, that's a design decision.

You're trying to tell me that's my fault, but it's not.

They design the driver to only bump fan speed to combat temps when it crosses that temp threshold, which obviously means that's well within card's specs.
(Or that they're retarded, which i already stated, you decide).

If that wasn't "normal", they would have designed the drivers differently.

Get it?

It's not that hard to understand really.

It is not a simple user's responsibility to modify cards to run safely at default conditions.

Which is why i say it's obviously a safe temp to be idling at.

And if not, they've overlooked something, & potentially they end up with an RMA (assuming that's too hot, which it obviously isn't).

And you still dont' seem to get it. Unless the ambient temps in your apartment are warmer than the ambient system temps around your video card, circulating air around the card will absolutely cause your card to run cooler. Removing the side of the case does very little to remedy this if there is no air flow to remove the pockets of hot air that build up around the card. In fact, removing the side of the case can actually result in higher temps for cases that have a side air blower, even though the ambient temps outside the case are the same.

You're assuming the 80C threshold means its running well within specs, which is probably the case in a perfect world. But assuming this is how it plays out in reality is plain ignorant given the myriad different hardware and software configurations out there. So, you can assume everything is going to run fine at 80C when everyone else is telling you that's hot and possibly problematic, or you can do what most others would when experiencing problems and try to remedy the situation. Until then, feel free to enjoy your melted ice cream (while pointing fingers and scratching your head).