can overclocking kill a GPU?

[RamonGTP]

I killed a GF4 Ti4400. i have two or three of my friends who have killed a 9800Pro. My cousin damaged a 7900GTX. It still worked but had to be UNDERclocked to be stable. None of the deaths were due to extreme OC's either. They were all operating at stock voltages and the 9800's actually had aftermarket arctic coolers installed on them, so GPU temps were well under control.

So yes OCing can kill a GPU.

 

[x3sphere]

I killed a GF4 Ti4400. i have two or three of my friends who have killed a 9800Pro. My cousin damaged a 7900GTX. It still worked but had to be UNDERclocked to be stable. None of the deaths were due to extreme OC's either. They were all operating at stock voltages and the 9800's actually had aftermarket arctic coolers installed on them, so GPU temps were well under control.

So yes OCing can kill a GPU.

Or the cards could have been faulty from the start. Increased load from the OC might've just gave way to the inevitable quicker. Unfortunately there's no real way to tell.

 

[RamonGTP]

Or the cards could have been faulty from the start. Increased load from the OC might've just gave way to the inevitable quicker. Unfortunately there's no real way to tell.

The answer is still "yes"

 

[BababooeyHTJ]

Obviously overclocking can kill components, just look at all of the early 4870s with degraded memory.

Like mentioned amps kill. Just don't oc the piss out of your card and run furmark fullscreen with the most stressful settings possible. Use something like GPU tool windowed to test the stability of your clocks and monitor temps for a while. Obviously do the same while gaming and you'll be fine. I noticed a lot of people blaming their clocks for taking out cards but do you have proof of that? Two 280s at the same time doesn't sound like a problem with clocks to me.

Or the cards could have been faulty from the start. Increased load from the OC might've just gave way to the inevitable quicker. Unfortunately there's no real way to tell.

exactly

Who knows if the higher clocks even made a difference?

 

[RamonGTP]

If amps kill then increased clocks make a difference. It's that simple. The debate should be whether or not it's an appreciable or neglagable difference, not if there is merely a difference, becuse their is. Incrased clocks = increased power draw.

 

[BababooeyHTJ]

If amps kill then increased clocks make a difference. It's that simple. The debate should be whether or not it's an appreciable or neglagable difference, not if there is merely a difference, becuse their is. Incrased clocks = increased power draw.

It's not that simple. Every architecture is going to act different. Thats why I mentioned the 4870 since a change in memory chips made a huge difference. Just do a little research to be sure that people aren't having problems and don't run furmark at the highest setting possible for long periods of time and you'll be fine. I could pull up a link that shows how furmark makes quite a few of last gen cards go above their tdp at stock speeds, or how OCCT with shut down any reference 4870/90/x2 at stock speeds.

I never run my cards on the brink of stability. I find out where that line is and pull back a bit.

 

[RamonGTP]

Obviously every card is different, that isn't the point and it is that simple. No one is saying you can't run into stability issues even at stock speed, so I'm not sure what your point is by bringing that up. So you can get a stock 4870 to shut down using furmark... That's great, but the answer to the question is still yes, overclocking can kill a gpu.

 

[BababooeyHTJ]

Obviously every card is different, that isn't the point and it is that simple. No one is saying you can't run into stability issues even at stock speed, so I'm not sure what your point is by bringing that up. So you can get a stock 4870 to shut down using furmark... That's great, but the answer to the question is still yes, overclocking can kill a gpu.

No, thats not necessarily true in all circumstances. Thats my point. A moderate overclock with good cooling, will not kill any card. Thats clearly just not true. Why do you think every "manufacturer" sells these overclocked cards that are no different than vanilla cards like Evga's FTW or XFX's XXX cards.

My point is don't be an idiot and overclock the piss out of your video card and run furmark. Hell don't even run long periods of furmark at stock speeds. Most importantly do some research before overclocking to be sure that there is no common problems.

 

[RamonGTP]

At the risk of repeating myself, the answer is still yes. I agree with what you're saying, but again, that wasn't the question asked.

CAN overclocking kill a GPU? Yes, it CAN kill it. Nobody is saying WILL overclocking kill a gpu, merely if it can. The OP had two cards that both had moderate overclockes on them and they died. Was it the OC? We don't know for sure. COULD it have been because of the oc? Yes, it could have.

To argue that is to say that it is beyond the relm of possibility, which it clearly is not.

 

[Ualdayan]

Okay well I've been wondering this for a while now...

I've got two GTX 285's and my first set died on me. They're EVGA FTW editions so they have a slight bump in volts..

You probably bought them at the same time right? Same batch? It's not really surprising they would fail at the same time when operated in the same environment. Personally I've had terrible luck with nVidia based cards since the 7 series - all you can do is buy from one of the manufacturers with a lifetime warranty (which you did) and RMA them over and over until you get tired of dealing with it or you get one that lasts. At least that's what I did with my old 7900GTX.

 

[BababooeyHTJ]

At the risk of repeating myself, the answer is still yes. I agree with what you're saying, but again, that wasn't the question asked.

CAN overclocking kill a GPU? Yes, it CAN kill it. Nobody is saying WILL overclocking kill a gpu, merely if it can. The OP had two cards that both had moderate overclockes on them and they died. Was it the OC? We don't know for sure. COULD it have been because of the oc? Yes, it could have.

To argue that is to say that it is beyond the relm of possibility, which it clearly is not.

I didn't see that 850mhz on a GTX285, it probably was the OC. There is just no way that was stable. Lol, I can only imagine. That is just an obscene overclock.

That is why I made the point to say do a little research. I would love to know what he had the shader clock at.

You probably bought them at the same time right? Same batch? It's not really surprising they would fail at the same time when operated in the same environment. Personally I've had terrible luck with nVidia based cards since the 7 series - all you can do is buy from one of the manufacturers with a lifetime warranty (which you did) and RMA them over and over until you get tired of dealing with it or you get one that lasts. At least that's what I did with my old 7900GTX.

Those are obscene clocks to be even attempting to run on a 285. It's just a reference, binned card. It's exactly the same as a vanilla 285, no bump in volts, nothing. He was trying to clock the things from 650ish to 850mhz. Thats nuts, a 30% oc on the stock cooler? Dosen't sound like a good idea to me, I bet that GPU tool would have shown artifacts immediately. How can you blame Nvidia or Evga?

Also you might want to take a look at ATI's RMA rates before bashing Nvidia there.

What type of water block was the OP using?

 

[XFreeRollerX]

You probably bought them at the same time right? Same batch? It's not really surprising they would fail at the same time when operated in the same environment. Personally I've had terrible luck with nVidia based cards since the 7 series - all you can do is buy from one of the manufacturers with a lifetime warranty (which you did) and RMA them over and over until you get tired of dealing with it or you get one that lasts. At least that's what I did with my old 7900GTX.

Yeah I got them at the same time...


The 2nd set I got (bought both at the same time, the serial numbers differed by only one digit, two counts down...IE: 100054 and 100056)

The next set was a lil more different, but they're lasting so far.