Dead 580m. Was using .92v vbios

Thought i would let you all know- my 580m died on me along with my motherboard. I was using the .92v. So any potential flashers, i suggest you dont run the .92v for long. I had been using it for about 3 weeks.

Maybe a mod could create a poll to see how many people have dead/problematic gpus while running .92v vbios. That will give a clearer picture about how safe it is.

 

Were you overclocked or stressing it much? I know you live in a hot climate. My first 580m also died running the 0.92v vbios. Who knows. I know many people using the 0.92v vbios and have been fine for many months so

 

Sorry for your loss, Chaos.

Mine didn't even last a day...

 

guys...I've been saying this over and over: You can overclock video cards all you want, but don't overvolt...it is a great way to destroy GPUs. If you want to run benches, overvolt temporarily, get your benchmark numbers, then set it bac to stock

Sorry to hear this Chaos

 

Widezu, no overclock, only overvolted.

Geocake- Sorry to hear that dude. You had overvolted the 6990m(in your spec) or the 580m?

Slickdude- yeah i learnt that the hard way!


To all others- I strongly suggest you all follow slickdude's advice. Dont overvolt for long.

By the way, if i want to overvolt it to more than .92v, how do I go about it?? (Oh common!! You guys weren't really expecting me to stop now, were you?? Us benchmarkers are an unreasonable lot )

 

Oh well, guess the hot climate didn't help either Sorry to hear.

 

What are you ambient temps, widezu?

 

Oh I live in a student flat so heating is non-existent for me. On a bad day, My ambient temps may never reach 15C.

How is your card dead? 8 beeps?

 

15c? you need to get a portable heater lol

 

I just got one. It was mainly due to the fact that my heating system has been kaput for a while and only got fixed yesterday.

 

Currently, my laptop doesnt switch on at all. On power/ batter anything. Motherboard is gone. AC adapter also being replaced. I was on .92v vbios and running furmark again and again.

My gpu was replaced some days earlier. Yeah i used to get beeps then but dont remember how many. This happened while I was on .92v too.



By the way, if i want to overvolt it to more than .92v, how do I go about it??

 

You can't 0.92v is the highest. And I reckon that Furmark killed your GPU. Why the need to be running Furmark so much is beyond me. That's your culprit right there.

 

You just shouldn't overvolt, in the past week we have had many people come in with dead 580m's after using the .92v bios. Plus, if you are not OCing then why even bother? All you are doing is making more heat and shortening the life for no reason at all. Just flash the .87v vBios that fixes the throttle and be happy.

Also, I would never use Furmark, it is a useless application that is consistently killing 580m's.

 

*weep*

 

I had repasted with pk-1 so was going through some heat cycles. Also was trying to find different positions for better cooling on my cryo lx.
Learnt my lesson though!

 

Are covered under warranty? You should be

 

Yeah I have warranty till 2016!! Thank god for that.

 

Your computer *IS* your heating system.

@Chaos, sorry for your misfortune.

 

Considering a stock 6990m runs at 1.1v I don't think it was the gpu that killed the motherboard. I think it was Furmark causing it to pull too many watts. Just a guess anyway. Probably crowd the mxm slot.

 

I agree with you Yoda, i learned the hard way, I used the .92v and my 580m went dead when overclocked very small amount after a few days. I now have stock voltage, and no problems.

 

Wow, that is a lot of 580M deaths. I asked in the vbios thread about the number of casualties and widezu said he was the only guy.

Gee, thanks guys. =/

My laptop did not switch on after the shutdown either. I had to remove the power cable and start on battery (about 5 minutes later after I recovered from the shock myself). Was then greeted with 5 beeps and upon further inspection the discrete NVIDIA card would not be detected in bios.

 

Man its like a masacre on alienware machines lately . I feel so bad for you guys

 

As far as I know it WAS only Widezu up until the last week or two. There have probably been a dozen people show up in the last few weeks to report dead 580m's caused by the .92v bios.

 

Yeah seems like everything was hunky dory. Mine died over New Years then everything was quiet over a month or so then boom. Like a plague.

 

With this and the Stop Ship order of the Cougar Point (Intel 6-Series) chipset that's in our R3s, what do you guys think is the likelihood that Dell/Alienware will replace returned R3s with the new R4s once they're out? I think I'm going to invest in an extended warranty and accidental damage.

 

I don't know. People say that Dell replace your entire system when you have 3 repairs on the same component. I'll put that to the test once the R4 comes out.

 

That stop ship thing is a year old, it was the SATA issues that all the new Sandy Bridge mobos had at that time. It has been fixed and is not longer relevant.

Honestly, I don't think this increases the chance. Basically, as long as Dell can get the parts then you are gonna get the same thing back in a repair, once the parts dry up then you will get the current parts. I would expect Dell to be able to get the current parts for at least another 1-2 years. However, if you get a 3 or 4 year warranty and get a failure later in your warranty you will very likely get an upgraded part.

 

Did Dell/AW fix it with a new stepping of the 6-Series chipset? How can one poll the chipset stepping and determine whether it's an old or new one?

 

Open cpu-z and click on the mainboard tab (or something similar i dont remember). If your is B3 which mostly everyone's should be then you are fine.

 

You have nothing to worry about. THose issues are a year old. If you laptop is new and it looks new, you have the latest chipset with the fixed port degradation issues.

 

It is B3. Thanks Long and Slick. Kinda wished I had a possible path to an R4 recall upgrade...

 

Got my motherboard and ac adpter replaced today.

Problem still persists. Laptop doesnt power on with either battery or ac adapter.

The blue light in the new ac adapter goes off the moment it is inserted into the laptop. So I am again getting a new motherboard. The technicians think it is an issue of reverse current.

 

that sounds all awful..Can't believe that many 580m went out like that...now I'm tempted not to use the .92v vbios anymore but in addition I don't have any warranty unlike all of you.

It still sounds weird to me, could anyone confirm they've been running Furemark for some while and think that it would be mainly the cause of the death ? sounds like a sad story to me...

Good luck with your replacing Chaos.
GeoCake, did you have any warranty with your 580m ?

 

Chaos, what warrenty do u have? Does a simple warranty include such stuff?

 

read entire thread and all i see is a couple of misfortunes and no clear reason. If its to be determined .92v is enough to destroy these gpus in gaming then a poll should have been used. Plus all those who had used furmark at all and had a gpu die at .92v should not be a statistic because that was likely a killer to gpu mxm slots and motherboard. these draw 100 watts normally using furmark and overvolting will inevitably lead to gpu death in a very short time guarantee the total draw power went well beyond that 100watts. Those who use .92v and stock clocks is not a big risk but strange to do because up to 745 core or whatever i heard .87v does just fine so stock clocks + .92v isnt dangerous but certainly pointless because as mentioned above it added heat and for no gain at all only reason we overvolt at all is to have stability and support of higher frequencies. Anyone want to report real results like Widezu... Whom actually had them die on .92v and no furmark?

 

I have lots of experience killing desktop vid cards by increasing voltage. I don't see why it is any different here with mobile cards. i don't think that it's a couple of unfortunate cases at all. From what i have seen, if you overvolt to .92v and you stress your 580m by overclocking and benching, you will eventually kill your card

 

in that respect i completely agree Slickdude and is kinda what i mean to say it seems those abusing or pushing cards to hard are having them die left and right at .92v but lets say you go 800core at .92v and just game... I see it being less of an issue... why people like you said expect a mobile card to survive the same or worse punishment a fullsize desktop gpu takes makes no sense. If .92v is used within reason it is not insta death for the average card... If furmark and numerous vantage and 3dmark benches are done over and over sure no doubt the voltage will be too much and more than likely because at .92v for each mhz .87v did the wattage only added to 100watts TDP... Im sure using .92v in a lot of cases is causing the gpu to be surpassing that TDP by a lot. I personally am going to light bench on .92v and call it quits after just a few have been done. Then i am going to set it to gaming clocks on .92v above what .87 handles (whatevr it was maybe 750 core max i dont know).
I will be careful and i wont wonder if mine dies why it happened but likely i wont be pushing it hard enough to do so. Not sure why people are pushing these cards that hard anyways what did it ever accomplish. I will give example of what it has and point out that this set of cards are not broken... Johnkss in the M18x Bench thread has some top 3d bench scores attained by pushing his card seemingly more than anyone else here... Using .92v and his cards are running just the same. Other than furmark he probably pushed them harder than those here who have had their gpus die based on the clocks and scores he used. Like 890mhz core on 3dmark11 at .92v i mean if that doesnt kill it what will. I know furmark is waste of time i used it on cars ive had and even had stability in it but on same overclock had probs in game. So its useless as a testing tool for normal usage. John keeps his cards cool as should anyone using an overvolt. Although increased voltage can kill a gpu and cpu the number one killer is actually the effects of not enough voltage. A gpu can die just as easily on stock volts with an overclock that requires too much voltage just like a cpu can. I learned this hard way undervolting 6970m's. 5870m's i had never had issue with higher volts. If the users dont keep appropriately cool they are shooting themselves in the foot because a throttling or overheated gpu is going to under perform a lower clocked stable one.
Flashing a vbios is unsafe in itself and playing with voltages is no different. Users who dont know what they are doing should not be flashing cards... even if cards are successfully flashed if overvolting you should have already improved upon stock cooling because these cards and heatsinks by design are only designed to wick away 100watts of heat. I know a lot of users on the overvolt just gaming and enjoying their system. I am going to do the same and those who dont want to take a risk shouldnt even attempt to use .92v overvolt... Because thats the risk taken by flashing or changing volts. So it does seem .92v can kill 580m's but not in just using them for games.
Slickdude i didnt write this to oppose your post but to reinforce that your right in the right scenario .92v will kill 580m and not being subjected to high stress its not as likely bot not impossible for the gpus to die. This is why a poll would help. We could see how many died at .92v at low clocks and how many dies at high plus out of them how many were pushed in an extreme sense and how many just died with as little as surfing the web? We need to say same for this... Flash responsibly.

 

Yes I had been running furmark repeatedly.

I have advanced warranty. But i think a simple warranty will also cover your graphics card for sure.

I was using .92v and stock clock for 3 weeks or so.

 

@Chaos I have to ask, why were you using the 0.92v bios at stock clocks? Kinda defeats the purpose and increases your heat needlessly.

 

Well i was benchmarking. 3d mark06. Trying different clocks but i would set it back to normal after tests. Just never flashed back to the original volt vbios.

Plus, for some reason, .92v throttled less in real life performance that the standard one.

Anyway, what's done is done. From next time, i'll keep th .92v for only a day or two in case i wanna benchmark then go that to the .87v one

 

Good idea Sorry I didn't know you were benching, now it makes sense

 

I agree sorry didnt know you were benching seems everyones on same page be very careful if using .92v vbios

 

Well I must say this thread has made me think twice about the .92v Vbios. Just wondering if a 580 without the .92v bios can keep up with a 6990 in Vantage, 3DMark11, 3DMark06

 

It is still possible. Plus when talking about benchmark scores, stability goes out of the window. To be a top dog, you need to choose either stability of extreme performance. If you want the most stable system, then you shouldn't have to worry about competing benchmark scores.

If you worry about benchmark scores, then you shouldn't be worrying about stability anyways.

 

been there my friend

btw sorry to hear that chaos, I hope you will get it somehow fixed (did you call dell yet?)

 

I had 6970m die on 8beeps so its nature of the beast we beat them up with benches we're bound to have problems some day... if your overvolting or not