did I damage my nvidia gpu by undervolting?

in 31st december 2009, my laptop vga was broken all the sudden. I was testing pcsx 2, the yellow/orange lines suddenly appeared in my monitor and my os was frozen, I did hard reset, but the yellow lines still appearing. the third reset, and my monitor was blank. it seems my vga was broken all of sudden, and I just knew nvidia 84xx and 86xx series are known for their defects. luckily I was still covered by warranty, the laptop just out yesterday and my data is safe.

my friend suggested that the damage probably was done because I undervolt my processor (not the GPU itself). My processor is initially 1.2500 v, and I reduced it to 0.95 v using rmclock. I was running stability test before, and I never encountered problem such as BSOD, so I thought it was safe to use my laptop in undervolted condition. the gpu was broken after 1 year 7 months.

but I'm curious, is it actually possible that I damaged my gpu by undervolting my processor?

 

Highly unlikely. Undervolting might result in a lockup because it doesn't get enough juice. It also wouldn't result in display issues, most likely the GPU or even the LCD.

 

thanks for the reply. my friend is working in IT department. he is more knowledgeable than me, but I dont think he is expert in hardware as well. I wish to keep undervolting this laptop because it saves power and reduce heat

anyway, I'm still waiting others for comment

 

nVIDIA® GeForce® 8600M GS

I think that speaks for itself as to what the problem is.

 

As Peon mentioned, your GPU's model number speaks for itself. A high number of those particular GPU's had serious problems and Nvidia admitted this. As for your friends suggestion, it is pure speculation.

Undervolting will not damage your CPU or GPU. The OEM's just go with a voltage high enough that is known to be near 100% stable across all chips. When we undervolt we need to spend a lot of time testing for stability which varies from one chip to another.

 

my 8600m gt did the same thing when it died but i had pink/yellow lines and clouds on my screen.

 

Ya...

I feel sorry for whoever your buddy does IT for...there's just no way a CPU undervolt could cause solder to break on that g84 core...in fact, it probably would have happened sooner without the undervolt.

 

I guess the verdict has been judged eh... I guess I must buy a new laptop before this happens again. sigh, this is really disappointing because I purchased this money for $929. I guess I better start looking for new ati-based laptop -_-

 

Any laptop without 8600m or 8400m GPU will suffice. nVidia corrected the issue with subsequent GPU releases.

 

Well, you've RMAed it, right? Don't write it off just yet.

 

Not really. The 8800m is also faulty. It's just not as bad as the 8400 and 8600.

 

btw how about the 9xxx series? I see some recent laptop with good price. I'm actually interested to buy one equipped with 9300 or GT130M

 

I'm not sure about the 9xxx series. But I think that the 9800M is a rebranded 8800m (that's what google tells me).

So I would stay away from that for now.

Go for ATI to be safe. Or get 3 year warranty.

 

the 9800m is a die shrunk version of the 8800m i think. i would say a 9800m is a safe purchase.

 

I haven't heard of any problems with the 9000 series.

Every high end Nvidia mobile card since the 8800m is a rebranded 8800m, including the GTX 260M and 280M. Ever wonder why they don't support DX 10.1?

 

I'm particularly interested to nvidia 9300 series or ATI 4570, there are two interesting laptop with pretty good price. I would love to go for nvidia, but this experience make me seriously switch to ATI for gaming until nvidia do something with their defective products.

 

The 4570 is way more powerfull than the 9300 and it's not going to die.

 

I would also recommend the 4570 as well, because the 9300m is not even a low-end gaming card. You won't exactly be maxing out games on the 4570 either, but at least they'll run at playable framerates.

 

ATi always give something cheap with great performance.
Nvidia always expensive. No doubt.

About TS's GPU, yes, it's the problem is created by the GPU itself.
Nvidia 8-Series Mobile GPUs = Problematic
Sad to say that, TS, your GPU is dying. RMA it if your warranty is still there.

 

I would get a new laptop... you can easily find nice cheap gaming laptops for $1k ish... good option is ASUS G line up .... G51J starts from $1499 or less or if u got less cash , u can get a ASUS G51VX with GTX260M.... or u'd want to wait a while for laptops with ATI's 5000 series and NVDIA mobile 300 series... the 9000 series is becomming old and hot....

 

Fingers crossed,knock on wood,horse shoe up etc etc...


So the "IT" guy thought under clocking the CPU killed the GPU.

Do they work for BB's Geek Squad? sure sounds like it :laugh:

 

I saw a pretty cheap toshiba laptop with ATI 4570, but unfortunately, it's dual-core, not core 2 duo, with 1 mb, is it still worth to buy?

 

Depends what chipset it has. Must be the P45 right?

If so then you could upgrade the CPU with a C2D for $100 on ebay.

 

If configured in "safe Vdc range" UV won't brick off your Hardware, I UVed my laptop, did the stress-test for 5 hrs, and it still Okay. we got similar problem, but my GPU is not totally died, Now i used it in standard VGA mode.
http://forum.notebookreview.com/showthread.php?t=448483

thanks for all sharing about 8xxx defect

 

Undervolting your CPU was hardly responsible for what happened with the gpu.
Undervolting results in a lower power drain, less heat output and longer lifespan of your laptop.
The worst thing that could happen is that you get a BSOD as a result of too much undervolting because the cpu is not receiving enough power, but once you find a stable low voltage that your cpu runs on under full load without problems, then there would be no dowsides really.

With all respect to the IT guy, but he has no idea what he's talking about.

What happened to your gpu is quite possibly a result of the soldering defect to which Nvidia admitted.
In any event, you sent your laptop for repairs under warranty and hopefully, your manufacturer will replace with with a non affected gpu.

Btw ... I would recommend you undervolt your gpu as well ... it will most definitely reduce the heat output even more and increase the longevity of your laptop (and quite likely the gpu).

 

Sorry to resurrect the post, but I had to defend OPs IT guy friend. I have the HP DV9500 with 8600M and have done EXTENSIVE research and testing including reflowing the board several times. I could rebuild the thing in my sleep by now. Anyway, the whole story is not included in the OPs summary of the IT guys diagnosis.
INDIRECTLY, undervolting the CPU is quite likely the cause of the failed GPU. In my laptop at least, the fan doesn't kick on until the CPU gets hot or GPU hits 80C. So undervolting and therefore lowering the temp of the CPU would put that much more of a difference in temperature between the two. I've seen my CPU running at 35C while my GPU was at 75C and climbing fast. Having the processing put onto the GPU instead of the CPU exacerbates the issue. ALSO, just overheating the GPU isn't always the cause of failure, the rapid change in temperature causing the solder to break loose is often the cause as well. If the ambient temperature around the laptop was cold and the CPU were undervolted and remaining cool, turning the laptop off while the GPU was over 100C could suddenly cause a change in temperature. So could running a graphics intensive program that left the processor and fan low but raised the GPU to 120+C.
ANYWAY, my point is that there's much more to the story than just "Undervolting my CPU killed my GPU." But I believe that indirectly that certainly could be the case. The timing with the undervolt and the seasonal temperature support the possibility as well. I learned these things the hard way, but hopefully it will be useful to someone else who winds up here by searching for info as I did. Take it for what it's worth.

 

The GPU you mentioned in your post is yet again, one of the known faulty ones from Nvidia.

In some cases, I do agree that (indirectly), undervolting the CPU could make the temperature difference between the two components larger and therby the potential of causing earlier or more damage to the GPU more likely and if the Bios that controled the fan was not very well implemented, then I guess it could be more of a GPU damage potential.

In the original post, if the IT person meant this, then they should have elaborated on their point.

 

Indeed that elaboration was required.
At the same time though, I don't think there were virtually almost any reports of gpu's reaching problematic temperatures because of their fans not kicking in due to the cpu operating on a lower voltages.

Then again, that's what temperature monitor software (free and resource light) is for.
They give you an indication if temperatures are imbalanced, so you can actually act on this.

This is why stress testing of cpu/gpu is performed so the new (lower) voltages can be examined properly as is temperature monitoring for the duration of the testing.

Bottom line is, it comes down to the user who is doing the procedure to make sure the hardware is working and hardware is properly cooled.

Any procedure performed on a hardware can hypothetically cause damage if you don't know what you're doing.
But if you do know what you are doing, and if you want to use new settings for the foreseeable time you are using the hardware, then stress testing and temperature monitoring for at least several hours in a row would be recommended (although the issue with temp imbalance would likely manifest itself relatively soon in the tests ... such as within the first 30 mins or so).

In this particular instance, the gpu was from a faulty series, so there is a very high probability it's premature 'death' was caused by the soldering defect and not undervolting.