this is the only community i feel i belong so i'm posting this here. i hope there are some pc hardware gurus around here. i'll take my chances.

for me it is a beast of a card and got it pretty cheap from a local store but in its default settings this card is burning hot!

dissassembled it to change the thermal paste and since i'm really inexperienced i wasn't sure if this cooler is in a great condition. notice how it looks like it's damaged right around that part where it is making contact with the gpu? is that really a problem? or is that really "the" problem? if so the card is already void from warranty because i unscrewed the cooler part, even though didn't cause any damage while doing it (that's how things work in my country).

changing the thermal paste didn't help so i downclocked it a bit to stabilize at a 90 degrees celcius. it wouldn't be a huge issue for me if the card wasn't throttling (fps drops) or making very loud noises trying to get the temps down.

AMD cards run hot but that thing shouldn't be running that hot. 90 degrees Celsius is ridiculous for a video card and running it like that will kill it fast and could even damage other components in your system. 70c is a more realistic temperature but I'm honestly not even comfortable with that.

I don't know what you're using but I'd personally use something like Arctic Silver 5 thermal paste. It's great stuff that most people swear by but it has a curing period so you won't notice a big change in temperature until it's had a few days to set in. The damage on the heat sync isn't the issue.

The overclock you have could be the culprit as could not applying thermal paste properly. Use a program like MSI Afterburner to get your overlock stable. Do not just push all the sliders to the right and think that it'll be ok. Every card is different and they all have their sweet spot. Having a massive overclock can kill your framerate as well. Also use that program to control your fan speed.

Your temperature issue could also be a whole slew of other things like not having antiquate air flow in your case. Just start small with applying some thermal paste properly and then controlling your fan speed and getting that overclock in check. If it's still running super hot then you may just have a faulty card.

Yeh, clean it up the best you can and apply some fresh thermal paste. It does look like an low quality block but not broken.

You should've brought it back. Cards shouldn't throttle back below standard clocks under load. If the new paste doesn't work, I guess you could downclock the card to standard 390 levels or you can replace the whole cooler with an aftermarket one if that doesn't help. Or try to return it.

This card should run 70-80C under load according to the reviews (290-390 cards run hot hot hot hot!).

i did change the thermal paste, with no effect. and yes, i use msi afterburner but it's not overclocked. only some minor factory-setting overclock. Arctic Silver 5 thermal paste is not available in istanbul, i used cooler master something something.

dropping the aux voltage and the voltage on the core from msi afterburner, reduced temps 3 to 5 degrees instantly and at %100 load the card is not throttling like it used to but i'm confused now, if it's not getting proper juice (voltage) how is it able to perform the same? or is it the same? because core clock and memory clock is stable.

setting up a custom fan curve from msi afterburner did not help either. unless i run it at a %70 fan speed, which is really loud, the card has 3 fans on it, it does not help that much.

i have a relatively good case nzxt s340 and i use 2x140mm fans to blow fresh air inside the case, 2x120mm to get the hot air out. so the airflow is supposed to be ok.

i only wanted to have a decent card that will run games at a stable framerate but this is what i got. but then again, i guess it's a lot better than the 650ti boost i used to have. with v-sync or framerate-capping the card is not running at %100 in games so i guess i'll be fine for a while.

thank you all for replying.

by the way gundogan. are you turkish? just curious.

@hkabaran: Half (dutch born), but I don't speak a word of it. :)

A lot of cards can be undervolted since most of them play it on the safe side, but like I said, the cooler should handle the card's stock settings.

@gundogan: oh i work for the dutch consulate in istanbul! i love dutch! =p

i don't want to do anything with turkish rma services for a few more frame rates so i guess i'll just learn to live with it. thank you guys for your responses. i'm declaring the card as faulty.

If you got it to be stable with lowering the voltage you are set to go. What happens very simplified is that the chip aint perfect an leaks voltage which makes the card fail its calculations when running in high speed. When that happens you get graphic glitches, chrashing drivers and in the end most likely entire systemcrashes and if you are really unlucky bricked cards/systems.

Sometimes increasing the voltage can offset the amount that is leaked so it runs ok but it comes at the cost of a higher wear and temperature which in turn can make the card destroy itself. Normaly the card throttles down before that happens which is your main problem problem.

If you can get it stable at lower voltage thats almost always a win but only if it really is stable. You might wanna do some extended full load tests to see its not failing and corrupting something for you.

It should be stable out of the box ofc. and this probably means you will have real poor results should you wanna try some overclocking but i can understand you dont wanna deal with RMA.

Edit: filled in with some more details.

What is your power supply? Amd cards run hot and take a lot of power. Also what is the temps of your mobo? There used to be a setting in catalyst for like a 20% power boost and if thats on it could be causing problems. Undo any overclocks even factory ones and finally could just be a bad card, and possibly why they sold it cheap.

i think it's this one thermaltake 550w. notice it does not directly support my 8pin+6pin connection on gpu by default. i use something like this to provide additional 8 pin power to the card.

factory clock speed of the gpu is 1050mhz. amd website tells me this card is 1000mhz. i lowered it but i'm done for the day. i will test it later and see if it makes any difference. catalyst settings seem ok as well.

i don't know the temps on mobo. i can only see them in bios but i guess that doesn't count right? they should be around (higher than) 40 celcius. but my room temp is also around 30 degrees (it is a hot summer).

don't get me wrong i finished a lot of games with this card in the past few months. i am ok with it. but it is loud and as the summer hits it is not getting any cooler. hope it survives the summer =).

@hkabaran: Yeah that PSU could be cutting it close at 550w i believe the recommended is like 600w? I know i skipped the AMD cards because i only have a 630w and didn't want to cut it too close. Also depends on your cpu, especially if your running amd as they also take up a bit of power plus all your other equipment take up power too. That COULD be causing the throttling at least, I say check the system temp to see if anything else is heating up during load. Hwmonitor is what i use(not the pro version).

Also what are your Idle temps? It's only hitting 90 under load right? I think the safest bet after trying everything is to return it the store and get another or RMA it with the company, taking the heat sink off could void the warranty though.

@oursin_360: An 80%+ efficiency 550w psu should be okay, unless that Thermaltake is just of poor quality in general or you also have one of those powerhungry fx9000 amd cpus. The adapter for the 8pin is not ideal, but if that's the problem then you should be having problems under any load I guess?

by the way my monitor just shut down for no reason! i had to reset the system and reupload this picture. this didn't happen before since 5 months i bought this card. i really hope this was a random driver crash or something. otherwise i will have to spend even more money on this pc. which i don't mind that much but still, not cool.

and also thanks everyone (again) for their insight. i learned quite a lot today.

The main thing to remember about thermal past on a GPU is you need a lot. Let me repate that - A LOT. Unlike CPUs wheere the functionla bit is under a heat spreader, a GPU on a graphics card is an exposed die. What that means is that the ENTIRE square of silcon needed to be covered by paste on a GPU. So when you put ist back together you want to apply enough of the paste so that will 110% cover that GPU die. Every sqare mm of it must have paste on it, done worry about putting too much too little is the fear.

All that damange you saw might have indiacted taht someone before you took the cooler off. If they did take ist off and then put it back on with too little paste that might have caused the issue.

As just make sure to apply the paste liberally, and give it a test.

@monkeyking1969: Just a sidenote. You absolutely want to cover the whole heatspreader with thermal paste when installing an CPU. Its not there for looks. Preinstalled fans and preapplied kits usually have an blob in the middle or an unnecessary thick "easy to use" paste which is why removing that and applying thermal paste of your own always is an good idea to improve the performance.

You want an even spread on the whole thing thats as thin as possible but it should clearly be covering it all. Putting on any more paste beyond that will just result in worse performance and a general mess on your card.

@gundogan:I don't like to take chances on it though, especially with any potential overclock. I say go with recommended unless you really know your shit, which i wouldn't say that i know the most about psu's lol.

@hkabaran said:

by the way my monitor just shut down for no reason! i had to reset the system and reupload this picture. this didn't happen before since 5 months i bought this card. i really hope this was a random driver crash or something. otherwise i will have to spend even more money on this pc. which i don't mind that much but still, not cool.

and also thanks everyone (again) for their insight. i learned quite a lot today.

I don't see much wrong there, try to reapply the thermal paste again like suggested and if it does nothing maybe the card is just faulty? Did you get a bluescreen when you crashed or did the screen just freeze? There is definitely something wrong

@oursin_360:just a black screen. i checked the cables and everyting but this dark grey lcd void stared at me for a while. then i hit reset.

i will buy a new thermal paste this week and update this if it works (or not) for future reference to other people. when i changed it i applied only a small amount to the center. didn't spread it all the way around the gpu.

considering buying a new psu as well.

it's funny because i chose this card over a 970 because it was cheaper at the time (and slightly faster) and i thought i wouldn't worry about low framerates for like 2 years with average graphics settings.

anyway i'll see if this new thermal paste and low voltage setting will hold off with mgs v and fallout 4 for the week.

edit: i have a confession to make. talking about a subject that you don't know well enough with a group of people who knows their poo is tense. i am somewhat afraid to be seen as an idiot i guess. this thread has reminded me what my inner fears are.

@hkabaran: Everyone starts somewhere, i don't know a whole lot but everything i learned was from asking questions and/or googling threads with questions like these. That's why I always try to answer if i can because i know it's never just the OP with a problem so googling would never work if no one ever answered lol.

For example, i have never changed a heatsink on a gpu. I would try what @rethla said about how to apply that thermal paste and see if that works, if you already have some i would try what you have even before your new stuff comes in and see if it helps.

When it comes to the reference 290x, they actually hit 94 c during most gaming, and even have to throttle back in order to not exceed it.
This strix cooler should be doing better though.

How long have you had the card for? has it always been like this?Looking at this review you should be around 70c, especially seeing as your other components seem to be pretty cool in idle on your screenshot.
As you can see, their computer eats about 400 watt at max load, and your PSU has about 500 Watt on the 12 volt rail, so that is not an additional concern.

If your case and other components have normal temps but your graphics card is this abnormally hot, I'd say you've either got a heatsink full of dust, or very poor contact between the cooler and the chip ( due to either loose screws/missing washers or bad thermal paste ).I personally run an Arctic accelero xtreme 3 on my 290X and it typically doesn't get hotter than 60c, the PCB should be the same as yours, unless asus has a custom layout on their strix cards.
One solution would be getting a hold of one of those, if you can.

I thought the method most people agreed on as best was a blob in the middle, especially on cpu's where the actual chip is just a small area ( usually in the center ) under the IHS.but covering the area on the side is probably not a bad thing.

I can only give my experience using a hot card, i'm currently using a GTX980ti but it's the MSI version, it idles in the high 50's and in demanding games gets up to around 85c, there are users out there with better cooling situations than me who have it run cooler but for the most part that seems to be normal for the card. Hardware makers seem to suggest that anything below 100 is safe but i'd start getting paranoid in the 90s, if i was in your position i'd take it back and try another card.
My friend got a 980 also recently but it's from EVGA, the temperatures are faaar lower at about 20c lower than mine in both idle and busy, so even within the same card there are fairly different margins.

Sorry i don't have any technical advice for you.

This might be a silly question but do you have some kind of custom fan curve for your card in afterburner? I'm guessing the fans are spinning up close to 90-100% with these temps but it's an easy thing to check.

Well the whole point of the heatspreader is to get a bigger contact area so while you can get sufficient cooling with just a blob you certainly get more efficient cooling by utilizing the whole area as it is ment to.

A blob in the middle is usually recommended becouse its an foolproof easy to explain method but work it out to an thin and even paste to the edges with an credit card or whatever you have at hand instead. Its just 1min extra work and not really hard. If you got another 2min you should clean all surfaces and the creditcard as well as you can beforehand.

For cpu's I use the credit card method to spread it on the heat sink, i don't spread it thick or anything more like a real thin dust cut and then i put the pea size in the middle on the cpu before i mount. I found that it gives me a better result than just the pea method, but not by all that much the difference is usually negligible maybe 1 or 2 degrees. Now for GPU's i would have no idea how to best do it.

I have an r9 290 and 90-95c is, unfortunately, an expectation for me.

i have an r9 390 and my temps even under load don't get that high. My only suggestion would be the same as the others. Reapply the thermal paste (a proper amount, more is not better here, like you need an uncooked grain of rice size to do the trick) then blow out any dust from the heatsinks/pipes and finally you can use a program like MSI afterburner to create a custom fan profile if you need to. my temps under load are 75C (tops).

Edit: My other question is with regard to your case. Is there sufficient air flow? Is the natural path of your airflow in your case blocked by the card itself because i remember that card is gigantic? are their adjacent cards taking up a slot that could be affecting the cards ability to move air around it to cool itself.

Just for reference this is the case i'm using, there's lots of vents and fans so it gets a lot of air flow, if you're using a plexiglass case with less ventilation, that could be part of the problem.

http://www.microcenter.com/product/441632/EVO_III_ATX_Mid-Tower_Computer_Case

Just other ideas. Good Luck duder!

Yeah covering all of it is best for sure, I get that logic, I've just been a bit clueless on how much more it helps than just covering the area above the actual chip.
The argument i've heard and seen for the pea method has always been lower risk of air bubbles from letting the cooler push the paste into place, instead of having the whole surface covered from the beginning.
I did use a creditcard in the past, but after i read about the pea method i just started using it, not because it was faster, just cause i swallowed the argument about the air bubbles and stuff, even though i didnt have any temp issues prior to switching.

Since GPU's don't have an IHS I'd assume that it is more important than on cpu's to get as good coverage as possible. msi goes a bit overboard tho

if you've got a reference card, they're "supposed" to be that hot.

i would buy a new thermal paste and update you guys on the details if i wasn't leaving work dead tired, coming home only to fall asleep.

i know, furmark is overkill. it puts a lot of stress on the card but the old 650ti boost used to be okay with this. not throttling, not overheating it was just doing its best. i did not change the thermal paste yet. with 3 fans on the card, fan speed over %50 is unbearable for me. %60 speed does not seem to make any difference either (still not below 90 degrees but extremely loud). by the way is the fan curve ok? is it stupid? i want it to be dead quiet while watching videos and stuff. i also noticed the psu fan not slowing down after the test, still humming.

edit2: i might be onto something. call me noob but i noticed the system fans were not spinning up according to (in accordance with?) gpu temps. they were following cpu temps, which are relatively cool all the time.

from bios i tried to spin them up a little faster (fans are not pwm, mobo doesn't support). i boosted the intake fans (auxfan1&2) from 800rpm to 900rpm and the exhaust to 1000rpm and gpu temps were 3-4 degrees cooler. if i can manage a strong positive airflow throgh the case this might the answer?

@hkabaran : check out how the new airflow does in games too.
When was the last time you cleaned the front dust filter on the case?

@hkabaran: try a progam called msi afterburner, i used to use it for my amd cards to set up a custom fan profile. I honestly forgot about amd not having good fan control.

Word of advice if you use that program dont do a staight line curve as it caused stutter for me whenever the fan speed changed (7950 card). Set the fan speed every maybe 5 degrees so its not constantly changing speeds

This card is known to run hot. 85-90C is actually very common for the 390. Be wary of some of the advice you've gotten so far, much of it is incorrect. For example, the new AMD RX480 runs at around 85C, components in video cards are designed to operate at very high temps, so 90C is not a huge concern. You've gotten good advice to reapply thermal paste and clean the heatsink/fans. I've noticed that your case fans are set to a very low RPM, you could try turning them up to at least 1000RPM to improve airflow. Open air heatsinks such as this one will see a decent benefit from it. Another thing to keep in mind is the temperature of your room has a huge impact.

@hkabaran said:

this is the only community i feel i belong so i'm posting this here. i hope there are some pc hardware gurus around here. i'll take my chances.

for me it is a beast of a card and got it pretty cheap from a local store but in its default settings this card is burning hot!

dissassembled it to change the thermal paste and since i'm really inexperienced i wasn't sure if this cooler is in a great condition. notice how it looks like it's damaged right around that part where it is making contact with the gpu? is that really a problem? or is that really "the" problem? if so the card is already void from warranty because i unscrewed the cooler part, even though didn't cause any damage while doing it (that's how things work in my country).

changing the thermal paste didn't help so i downclocked it a bit to stabilize at a 90 degrees celcius. it wouldn't be a huge issue for me if the card wasn't throttling (fps drops) or making very loud noises trying to get the temps down.

sorry for not updating but i have some valid excuses like the attempted coup in my country (there was a heavy military mobilization especially in istanbul, where i live. now replaced by even more armed police officers. searching every bag in metro stations etc.) and then government's expected and "justified" oppression in every area... i feel soo down. at some point i didn't even care about the pc or video games. i tried to quit my job as it became too tiring for me but my supervisor convinced me to stay (offered less workload).

and the r9 390 of mine is more of a disaster (ok not so much). i managed to dismantle the cooler but failed to reinstall it. damaged some parts of the pcb. pc didn't boot with it and the fans were spinning %100 and then i ordered a 1060. from msi this time.

that's about it.