Should I RMA my 4870X2?

[cokenbeer]

So I've come to the conclusion that SOMETHING is wrong with my graphics card. One of the cores get up to 30 degrees hotter than the other, to the point that my system shuts down when I play games. If I sit at the desktop one core idles at 45 degrees and the other in the high 70's low 80's with a 30 - 35% fan speed. The card is NOT overclocked at all. I don't know if it might be an issue with the contact on the heatsink, or maybe some other issue, but I'm wondering if I should just send it back.

 

[RME]

I am not sure, but i heard that the card only loads one core if the second is not needed...
But i could have misunderstood

 

[emp]

Assuming you have proper airflow, start by increasing fan speed to 100% to test it, if it keeps giving you problems then definitely RMA, if not I'd say drop the fan speed in small increments until it starts giving you issues, if you deem it acceptable then keep the card.

 

[Nik_I]

before you send it back, try manually adjusting the fan speeds. i haven't heard of the 4870x2 having the same heat problems as the 4850, but it's worth a try. after all it's alot easier to try that than to have to RMA the card.

 

[JAYDEEJOHN]

Sitting at DT, its normal for one core hotter than the other, as only 1 is used in 2D. You say it actually blue/black screens? Or only when you play games? Whats the temps when you run games?

 

[nukemaster]

If it is crashing and all your other components are fine(PSU and such) I would RMA it. The card should not need any fan tweaks, i do not even use those on my 4870. I did try them, but the card keeps it self in the low 80's and that is about the same as my 8800GTX was at load anyway.

 

[No1sFanboy]

There are two reasons your card will run hotter on one core. In one of your other threads I linked you to a thread I had where I described how I've noticed one core running at full speed on startup. This would happen up until I had been into and exited a 3-d app. The card will also run one core at higher speed when a 3d app is minimized or in windowed mode. Currently this card does not utilize both cores when in windowed mode which I can't recall where but I think Anandtech pointed this out in their review.

The other reason the card one core will run hotter is because it receives air after the first core.

Temps of 75-80 are probably not a cause of your computer shutting down. If a card was over heating I would expect it to be throttling back and or artifacting before the whole computer shuts down. Being a new build you need to as strangestranger pointed out in one of your other threads prove out the rest of the system. Memtest and prime95 or Intel burntest.

That said it could be the card, do some sort of artifact test like atitool or furmark.

 

[cokenbeer]

I understand that one core will load itself if the other isn't necessary. But should one core be at close to 80 degrees? In a case with great airflow where all other components are cool (PSU isn't even turning on the fan, CPU stressed is at about 41-42 degrees, fans are exhausting warmed air and several intake fans are taking cold air from the top all in a cosmos S, and ambient temp is about 20 degrees) why would a single core on my graphics card be acting so strange?

While playing Spore earlier, a game that's not too difficult to run, I had the pc crash on me TWICE with my fan set at 45%. Both times by the time it rebooted the temps were reading in the mid 90's on the one core and 50 on the other.

I would prefer NOT to remove the heatsink unless I'm confident that there is not some serious flaw with the card. If I was certain that it's just a problem with the heatsinks contact then I would reseat it and be done with it.

As for RMA procedure, how does it work with NCIX? I'm still waiting on my second card to be delivered and I don't want to send this one back and have to wait until I can use my PC again.

 

[cokenbeer]

There are two reasons your card will run hotter on one core. In one of your other threads I linked you to a thread I had where I described how I've noticed one core running at full speed on startup. This would happen up until I had been into and exited a 3-d app. The card will also run one core at higher speed when a 3d app is minimized or in windowed mode. Currently this card does not utilize both cores when in windowed mode which I can't recall where but I think Anandtech pointed this out in their review.

The other reason the card one core will run hotter is because it receives air after the first core.

Temps of 75-80 are probably not a cause of your computer shutting down. If a card was over heating I would expect it to be throttling back and or artifacting before the whole computer shuts down. Being a new build you need to as strangestranger pointed out in one of your other threads prove out the rest of the system. Memtest and prime95 or Intel burntest.

That said it could be the card, do some sort of artifact test like atitool or furmark.

All components have been tested and seem to be working fine. I have my PC at stock clock speeds and my memory (DDR3 1600) is set at 1333 to match the FSB. Memtest ran without error and Prime 95 also ran without error and the CPU stayed relatively frosty!

Also after the comp shuts down and I restart the card temps are still in the mid to high 90's on the one core even after having a fan running for a about 30 seconds.

 

[No1sFanboy]

If you're still seeing over 90c after all the time it takes to restart then maybe you are getting some crazy high temps. Have you run something like GPU-z in the background to see what your temps were at rather what they are after exiting. In the time it takes to exit or at tab out of a game your temps will drop around 10c.

If you run GPUz in the background with "continue refreshing..." checked you can go back and hover the mouse over the high points on the graphs to see what your temps were.

 

[cokenbeer]

If you're still seeing over 90c after all the time it takes to restart then maybe you are getting some crazy high temps. Have you run something like GPU-z in the background to see what your temps were at rather what they are after exiting. In the time it takes to exit or at tab out of a game your temps will drop around 10c.

If you run GPUz in the background with "continue refreshing..." checked you can go back and hover the mouse over the high points on the graphs to see what your temps were.

Yeah it was hitting the 100 degree mark and then shutting down. The other core doesn't even go above 70! Also the chip in the middle (PLX?) is getting VERY hot. if I hold my hand over the back of the card I can feel the heat radiating off of the rear (near the vent) gpu and off the chip in the middle. The other gpu I can actually touch the backplate and it's relatively cool. The fan is moving a lot of air so that's not the problem.

 

[cokenbeer]

ddr3 at 1333 is not matching the FSB at 1333, it is twice as much. just a wee point, nothing to do with your problem but just a wee point.

Tis' what I meant

My CPU isn't overclocked so my ram is sitting at 1333.

I'm waiting for notification from koolance or danger den. I'm gonna pick up two water blocks and cool the card. My other card came in today and I put it in place of my old one...

Temps for the new one were high at first. Once I opened a 3D app and closed it I'm reading 46 and 39 degrees which is MUCH better! So my old card is being RMA'd.