Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Quick Links
The life span of a graphic card
I bought my Alienware Aurora ALX okay okay I know overpriced and all that in 2010. The card on the machine was Geforce 480 GTX which died in July 2012. I got a replacement from Dell since it was still on guarantee in august 2012 a Geforce 580 GTX,I think they did not have a 480 so they gave me an upgrade for free. That promptly died in a month or so and in November 2012 they replaced the card with a Geforce 580 GTX. The card is dying again this week. I think it is the card because the same cycle of events are occurring with display driver failing and freezing or recovering depending on my luck when I try to play a game. The Dell guarantee has expired though after 3 years.
Also when I ran diagnostic
Intel(R) 5520/5500/X58 I/O Hub PCI Express Root Port 3 - 340A PCI Express Status Test failed it says.
So that is about 15 months. Is this normal or is it that the cards that Dell have in stock have short life spans. I am thinking of asking them to replace it and may be upgrading to a GeForce 680 GTX not even sure if my machine can take that although the PSU is 850 watts but they have not updated bios on the motherboard so I think a 7 series is out of the question but am worried about spending too much on a 4 year old machine . It is a beast though cannot even lift it so damn heavy. The thing if I bought it and replaced it and something went wrong I would not have a guarantee to fall back on but if I buy the card from them and they replace it I can always go back and get them to come if something goes wrong.I paid a lot for this machine but it is still very good.
Just wondering about this short life span of the card.
Comments
Sounds like you got some bad cards somehow. Every nVidia card I have had lasted longer than the usefulness of the card. I.E 5+ years in most cases to the point where I had to replace it to meet the demands of a game, not because the card itself was failing.
Are you sure it's not a faulty motherboard/PCI E slot?
Currently Playing: ESO and FFXIV
Have played: You name it
If you mention rose tinted glasses, you better be referring to Mitch Hedberg.
that happens alot when you buy from alienware
next time build your own PC
Typically, a video card is designed to run for over 5 years. This falls in line with the typical upgrade every 3-4 years. However the Fermi based cards are a bit special. Quite simply they have always been unstable and it was estimated they would die after a couple years due to the excessive amounts of heat they output.
Onto your actual issue. There are 2 things that could be happening with the video card. A) Its a dud, there are always duds and these account for maybe 0.25% of cards from a manufacturer.
Its refurbished, Sometimes the issue seen is that something like lets say a capacitor went out where replacing it with a 25 cent replacement would fix it. Refurbished parts are typically good to go again for another several years since afterwards they are put through the same level of testing. However, I think they will have a much higher failure rate than a brand new card.
There are other things it could be. Dust, but considering its a replacement card that's probably not it. The PSU is lacking power for these cards. The GTX 580 is a pretty hefty card and if the PSU has degraded to the point it can no longer provide that power stably than it could also be the issue. Considering its a DELL machine they matched it with just enough wattage and used a low to mid-range PSU.
Power supply is 875 watts not 850 sorry .That is not enough for the 580GTX ? So if I get a replacement I should not try for the Geforce 680 GTX then ?
Also you think they gave me a refurbished card. No wonder they gave me an upgrade. Hmm can I trust them by buying a new card from them and getting them to fit it for me. I have changed graphic cards before but if anything goes wrong I have no one to blame. This way I can get a guarantee I hope so I can get them to come again to solve it.
The PC guarantee expired in 2013 so I have to pay now when they come.
Bumpgate. NV screwed the pooch on both 65nm and 40nm product lines. Buying one is like Russian roulette with 2 bullets.
680 is 28nm. no more problems. I would go with that.
It could be your power supply or maybe its your motherboard or you're just really unlucky! Iv had my 7870 for a year and it runs very hot being overclocked, while its lifespan is already a year in and no problems so far. I'm not a expert but maybe your motherboard/power supply is killing off your cards?
A bad power-supply can lead to bad things, maybe your power supply is supplying far less power than the card requires, or something just isn't clicking with the motherboard you have the card in. I can only give speculation. Best of Luck!
Just go for the GTX 680 it uses less power than the GTX 580. http://www.anandtech.com/show/5699/nvidia-geforce-gtx-680-review/19
<InvalidTag type="text/javascript" src="http://www.gamebreaker.tv/cce/e.js"></script><div class="cce_pane" content-slug="which-world-of-warcraft-villain-are-you" ctype="quiz" d="http://www.gamebreaker.tv"></div>;
Thanks the 680 gtx it is.
If it was the power supply wouldn't other things also go wrong ?
If the power supply has 2 rails and 2 6 pin connectors you can try the other connector since it would be isolated from the other rail. Just mark the connector that was being used and try the other one.
If the PSU primary were failing you would have seen instability immediately after the 1st replacment.
How old is the power supply? Model #?
Did you clean your cards often? Those 2 you had can fail at lower temperatures than normal due to soft fill material.
It is not normal unless you overclock severely.
Either the PSU or your electricity from the grid is spiking killing the cards. A good PSU would prevent this but Dell uses cheap crap.
There is of course the possibility that your something else causes the problem, like extreme heat, loads of dust or cat hairs or ESD (electro static discharges). You can feel if it heats up a lot by just sticking in your hand inside the case (or where the fans blow out air), if it feels more than a few degrees over room temperature you should consider getting more fans or a new case. Dirt or ESDs is unlikely unless you have several cats or open up the box and touches components.
There are ways to test the output of the PSU including software to see if the electricity is stable or not but if the spikes are rare it might take long time to notice them
Of course you might have very bad luck as well, graphicscards that last a month are not common but I seen a few with factory made failures that just died in perfectly working machines and the first 3 years is not that impossible even if most cards at least live 5 years.
Still a nice Corsair pro 750W or maybe an OCZ would not hurt, they actually saves the electricity bill since they give you the power you need but scale down when you don't need all of it, unlike cheap PSUs that always run at full power. The bad news is that some of Dells cases might need a little "holistic engineering" to fit a new PSU since they are custom made.
disreguard
Lol ninja edit
GPU's are not durable. They break easy. ESD is deadly to them. And they have very little overvoltage protection. The components are not designed to run out of spec.
Go look up Bumpgate. it affects both those cards he had.
I did some fixing on one of Dells newer custommade high end Alienware computers 2 months back, not that impressed. Beside that the whole thing were custommade with some rather different solutions then other cases (it did look really good though and that is just bad if you want to replace parts with different stuff) I was not impressed by a few things.
The issue with that computer was actually that it was raided and only had 2 settings: all drives in the raid or raid off, and the owner had put a new harddrive in a slot which killed of his raid and he lost his girlfriends precious pictures of her dead father. Took a little work to restoire them and the whole thing seemed like a weird solution to me, all raided computers I seen actually lets you choose which drives to add and which you leave as normal drives and have normal as standard if you put in a new.
But what I said was not that Dell have bad engeineers, I said that their PSUs are cheap (at least compared what I use). And that adding a new might mean that you need to make new holes for screws or more room, that doesn't go for all cases but some.
In my own case I had a Dell in the late 90s but custommake all my computers nowadays, I know exactly what to use my computers for which means I can design it exactly the way I want and my nice Corsair pro 950W have a very low malfunction frequency.
If OP post what type of PSU Dell put in his computer we could tell more though but the PSU is something almost all brand computers cheap out on, you can't see it on the speccs and a good PSU cost about twice what a cheaper do. In many cases do a cheap PSU work excellent (with the exception of higher power consumption) but in other cases it can give you a bunch of different problems, like sometimes rebooting, starting to sound a lot due to a bad fan or even killing of the motherbord or GFX card.
I wouldn't say easily except with ESD but it does happen, particularly when the PSU have a bad protection against overvoltage.
So yes, it might be bad luck but with 2 cards the odds are very high that the problem is elsewhere. Electricity spikes and heat are the 2 most common reasons that they break when they are in a closed case.
Computer parts are cheaper now then 20 years ago and that means that they also test them less, this is even more apparent in motherbords which have a far too high failure rate nowadays but it still is very rare that someone gets several bad card that isn't of the exact same model (if you have several of the same that could mean that the guys who made it had issues for a few days but still shipped, have happened with all parts from most companies. I had to replace 4 ASUS MBs of the same modell for different friends the same moth a few years back, hardly a coincident).
No, I think OP is worrying with good reasons.
It depends on a lot of factors.
One is power delivery. This means both from the power supply to the video card, and also as handled by the VRMs and delivered to the GPU and video memory. A voltage that is stable as current fluctuates and doesn't have much ripple puts a lot less pressure on the components receiving it.
Another factor is heat, both of the GPU and various other components on the video card such as the VRMs. More heat tends to make electrical components perform worse, and puts a lot more wear on the card. Heat is greatly affected both by how well the card is at getting heat off the card, and also by how good the case is at getting heat out of the case and keeping the ambient case temperature down. The temperature of the room that the case is placed in matters, too, though this usually isn't a big deal outside of some relatively extreme situations.
Another factor is build quality. Higher quality fan(s), higher quality capacitors and VRMs, better quality control on laying out components on the PCB, and so forth all reduce the probability of various things going wrong and killing the card.
Overclocking is a big factor. Increasing the clock speed of your video card adds considerable extra stress to it. If you increase the voltage as well, that's really asking for trouble. Lower clock speeds and lower voltages put less stress on everything and make it less likely that the card fries.
And then there's dumb luck. If card A has a 5% chance of frying within 3 years and card B has a 30% chance of frying within 3 years and you have card A, that doesn't guarantee that it won't die inside of a year. GPU chips are not the same at an atomic level, and flaws on that scale may well be imperceptible until the card dies after you've had it for a year or two--at which point, you still don't know exactly why it died, but only that something went wrong.
Buying from Dell, you don't expect a system to be great on power delivery, heat, or build quality. Or at least you shouldn't. They'll typically avoid completely awful on all three of those, but higher quality costs money and their customers won't know the difference outside of the binary outcome of it died or it didn't.
-----
The GTX 480 had a pretty severe case of overheating. The stock cooler on the reference card simply couldn't handle the card, leaving it very, very prone to overheating. You could kind of handle it with fairly extreme measures, but the odds that Dell would take such measures are basically zilch. The GTX 470 was similarly troubled, but other cards from that generation were merely inefficient, not highly prone to overheating.
It's not at all surprising that the GTX 580 you got from Dell was refurbished. Warranty service doesn't require them to give you a brand new card; it only requires them to replace a card that doesn't work with one that does. That's not unique to Dell, either. Some people send parts back for warranty service, Dell tests them out, and decides that the part works fine. So they send it off to someone else who wants warranty service. But that does mean that you're getting a card that someone else didn't want, possibly because he thought it was defective or maybe simply because he bought a computer and then returned it.
Sounds like the posters have covered the bases on this.
Heat is a real killer of electronics and GPUs are especially vulnerable. I try to keep my house at 70-75 degrees.
The quality of power. Everything from spikes to other equipment on the circuit can affect your system. I use a UPS on all my computers and sat equipment. I also make sure there isn't anything on the circuit that can problems like anything that has heater in it. I use 1000 Watt power supplies for my systems as adding another bit of protection.
It is unlikely that you have a ground loop current problem going on but just in case, make sure all items plugged into you system are on the same circuit and ideally from the ups. GLCs can cause weird problems.
Static electricity and electromagnetic factors can damage chips. Make sure if you handle internal components that you are using proper proceedures. They can blow chips or degrade them. At least with the blown chips you know right away while degraded circuits are worse in the long run.
Epic Music: https://www.youtube.com/watch?v=vAigCvelkhQ&list=PLo9FRw1AkDuQLEz7Gvvaz3ideB2NpFtT1
https://archive.org/details/softwarelibrary_msdos?&sort=-downloads&page=1
Kyleran: "Now there's the real trick, learning to accept and enjoy a game for what it offers rather than pass on what might be a great playing experience because it lacks a few features you prefer."
John Henry Newman: "A man would do nothing if he waited until he could do it so well that no one could find fault."
FreddyNoNose: "A good game needs no defense; a bad game has no defense." "Easily digested content is just as easily forgotten."
LacedOpium: "So the question that begs to be asked is, if you are not interested in the game mechanics that define the MMORPG genre, then why are you playing an MMORPG?"
Everyone has covered the "why" sufficiently.
However, if you haven't already purchased new parts for replacement, go to dell's website, punch in the service tag, and see if you can extend the warranty (and obviously, the price of doing so). Dell will allow you to extend those warranties for quite some time, and the price is usually reasonable (maybe, 200 bucks for another year). If the price is reasonable, you can extend it, then just have dell replace things until your system is back up and running.
Another method, if you want to, is to see if you can extend the warranty and add accidental damage to it (yes, you can get that on desktops). If you can do that (and if the price is reasonable), then take your system and drop it so that there is chassis damage. Once you've dented the chassis, call dell, tell them you dropped it going to a lan party or something, and they will build you an entirely new system. It's probably not well known, but with Dell warranties, if there is any chassis damage, they will not fix the machine, only replace it. With a machine at your age, they will replace it with a comparable CURRENT system (so, if you bought your machine with a 480gtx which was top end back then, the new one would probably have a 780gtx). So, if you go this way, you end up with a nice upgrade, at a fairly low price.
I just throw that out there in case you were not aware those are options.