Average lifetime of your discrete video cards?

[eccerr0r]

Seems like people do avoid used video cards like the plague. Curious how long your video cards last.

I suppose I specifically am curious about the discrete cards used for games or mining or something more than Word or Firefox. If you have a low end card but use it for games or mining, include it!

If you had more than one or worked with more than one, choose the average. It's your choice whether you pick mean, median, or mode, so don't take poll results seriously as this will skew things.

If you repaired it, count only the time to the first failure.

Knock on wood, I have my first GPU (Radeon HD 5770) that hasn't bought the farm *yet*...Not sure if it's because it's a mobile chip with its higher heat rating on a PCIe card and huge heatsink. All my other GPUs (Radeon, Nvidia) lasted like two years before they started crashing/artifacting.

If you ended up disposing of a card due to obsolescence, it's up to you to consider it a 'failure' to run the software you want to run or not

[mockingbird]

Re: Average lifetime of your discrete video cards?

A VERY good question...

...but very difficult to answer.

GeForce 600 and up have some very good new idling features. Geforce 1000+ is already 15nm or less.

So time will tell with the new stuff.

[Vishera]

Re: Average lifetime of your discrete video cards?

Knock on wood, I have my first GPU (Radeon HD 5770) that hasn't bought the farm *yet*...Not sure if it's because it's a mobile chip with its higher heat rating on a PCIe card and huge heatsink.

I lost some 8400GS cards during the capacitor plague, but since then no failures.

My HD 5750 and HD 7750 are still going strong. The last gaming duty that the 5750 saw was last spring, it's since been demoted to HTPC use. For something that's not worth $30, it's pretty great.

The unfortunate thing about buying used cards is that usually the warranty doesn't transfer to a second owner. EVGA and MSI (maybe Gigabyte?) do allow second owners to submit warranty claims, other brands do not.

[Vishera]

Re: Average lifetime of your discrete video cards?

Knock on wood, I have my first GPU (Radeon HD 5770) that hasn't bought the farm *yet*...Not sure if it's because it's a mobile chip with its higher heat rating on a PCIe card and huge heatsink.

I lost some 8400GS cards during the capacitor plague, but since then no failures.

My HD 5750 and HD 7750 are still going strong. The last gaming duty that the 5750 saw was last spring, it's since been demoted to HTPC use. For something that's not worth $30, it's pretty great.

The unfortunate thing about buying used cards is that usually the warranty doesn't transfer to a second owner. EVGA and MSI (maybe Gigabyte?) do allow second owners to submit warranty claims, other brands do not.

[mockingbird]

Re: Average lifetime of your discrete video cards?

My HD 5750 and HD 7750 are still going strong.

I am actually waiting for my 7770 to die because I have a lifetime warranty on it... Maybe I should leave it on in linux or something doing something intensive 24/7.

[brethin]

Re: Average lifetime of your discrete video cards?

Had a few fail but they were all XFX with a lifetime warranty so they were replaced free and most with better cards.

[eccerr0r]

Re: Average lifetime of your discrete video cards?

The unfortunate thing about buying used cards is that usually the warranty doesn't transfer to a second owner.

Well, if they never fail, why would you need to worry about transfer?

Seems that they do fail more frequently than, say a cpu or motherboard?

I had a GeForce4 MX420 fail and a RadeonHD 3650 fail on me. Both may have been heatsinking problems where the fan failed but nevertheless succumbed to failure even if the fan was repaired.

[brethin]

Re: Average lifetime of your discrete video cards?

Well, if they never fail, why would you need to worry about transfer?

Seems that they do fail more frequently than, say a cpu or motherboard?

I had a GeForce4 MX420 fail and a RadeonHD 3650 fail on me. Both may have been heatsinking problems where the fan failed but nevertheless succumbed to failure even if the fan was repaired.

XFX has, or had not sure they offer it now, a double lifetime warranty. So the second owner is covered as well if the orginal owner registered the card. I sold a few cards and the warranty transferred.

[mockingbird]

Re: Average lifetime of your discrete video cards?

Had a few fail but they were all XFX with a lifetime warranty so they were replaced free and most with better cards.

Can you please elaborate on what you got in exchange for what?

Thanks

[momaka]

Re: Average lifetime of your discrete video cards?

Hmmmm..... hard to give an answer to this poll. Title says "lifetime of your discrete video cards", but poll has an option "Only had integrated, never had failure". Seems like it's pulling in two different directions.

In any case, I've kept all of my "main" PCs with either old, low-end discrete video cards (and I really mean old, as in early/mid 2000's hardware) or onboard video. With those, I've never had any failures.

As for discrete graphics cards, I've had just about every one fail, unless I modded it ahead of time with better cooling and unless it was a really old low-end card. That said, it is worth noting that I've only had TWO new/unused video cards, and these were:
- a S3 Savage 4, 8 MB, 4x AGP. This was back with our Pentium II PC, which my family bought in 1999 and we used it until about late 2003, I think. I then retired that PC and moved it to my uncle's house, where it saw only a little bit of use in the summer for the next two years. The video card in that PC never failed, despite running hot enough to cause blisters if you happened to touch its heatsink. I did plenty of video gaming on it... or as much as one can do on a S3 Savage 4 video card anyways. I still have that video card.
- an ASUS Radeon 9200 SE, 64 MB, AGP 8x. This one, I bought with a custom PC I assembled after retiring the Pentium II PC mentioned above. I used that Radeon 9200 SE video card for about a year with its stock heatsink. Eventually, I found that the CPU ran rather hot due to case constraints, so I started adding fans in the PC and put one on to blow air across the video card as well. That Radeon 9200 SE video card was used with that PC from late 2003 to about late 2006 or early 2007 when the PowMax PSU in that PC developed bad caps (please don't judge me for my PSU choice - I was a PC noob back then, and even plenty of knowledgeable folks used PowMax, it seems... until the word about badcaps and their crappy PSUs spread out). I don't know if my added cooling helped or not, but the card still works today. And I did a ton of gaming on it back then, since I was in high school and had much more free time. The average per day was 1 to 3 hours of gaming, followed by an hour or so of homework. If anything, I think I've clocked the most hours gaming on that old beast. Between Half-Life Deathmatch, Half-Life 2 + Deathmatch, Counter-Strike 1.6 (and later Source), Garry's Mod, Need For Speed (4, 5, 6, 7, and 8), CMR3, and Mafia, it seems like there was never any downtime for that video card. Weekends were the worst for it, especially if we had bad weather - I could spend up to 8 hours on "gaming sessions" with very little breaks.

That said, I'm inclined to think my Radeon 9200 SE wasn't the only one to survive that long. If anything, you can find these old video cards on eBay by the bucket and I've never seen one advertised as failed or with artifacts. And I think that applies to a lot of early ATI video cards that were made prior to the flip-chip era (i.e. ATI Rage varaiants, Radeon 7000, 7200, 7500, 8500, and 9000). Until the introduction of Khan architecture (flip-chip design, which was used for the R300, and later R350/R360), failure due to artifacts was almost unheard of. Well, at least for ATI. On the nVidia side of things, I have seen GeForce 4 TI video cards (TI 4200, 4400, 4600) fail with artifacts, and occasionally some of their MX4xx and MX4000. But GeForce 3? Never. (Part of the reason why you'll probably never see an original Xbox failed due to a bad GPU).

And nowadays?... I have a small collections of all sorts of video cards - over 90% of which I bought used on eBay. Most are PCI-E, but I also have quite a few AGP ones, and several decent PCI ones. From the eBay ones, a good 70 or 80% I bought listed as "for parts/repair" - some due to bad caps, some due to artifacts, and some because they were just untested. The ones with bad caps were over 90% nVidia, if not 100%. I don't recall actually having to recap an ATI video card. But as for artifacts, that's a completely different story. Here, it's hard to say who has done worse, as I have a ton of failed ATI and nVidia cards - mostly high end ones. But the distribution is weird.

For ATI, I have mostly their mid/high end ones fail: namely a good stack of Radeon 9700, one 9800 XL, and also a ton of HD4850. All of these appear to have failed due to inadequate cooling. I also have a few mid-range and upper-low-range ones that have failed, namely: X1650 AGP, X1550, and HD6670. Again, all of these failed due to overheating (with the X1650 and HD6670 probably failed due to bad fans... the X1550 was a passive-cooled one - so I'm not surprised it failed).

On the nVidia side of things, I have failures from all ranges. Their low and mid-range cards were mostly plagued by bad caps and undersized coolers (though the bad caps seem to have failed before the cards died from excessive heat cycling). And the high-end ones were mostly due to inadequate cooling. I can post a list with details later, if anyone is interested. But literally almost every nVidia card I have has either failed or bound to have failed if I didn't mod it. For those of you who have seen my XFX GeForce 8600 GT and 6800 XT threads, you know what kind of modding I do. On that note, I'd like to issue a warning against all old XFX nVidia cards, as it seems quite a few of them had crappy GSC/Evercon,Sacon FZ caps. They finally stopped using that crap cap maker around the GeForce 200 series. Anything before that, be careful.

[momaka]

Re: Average lifetime of your discrete video cards?

Knock on wood, I have my first GPU (Radeon HD 5770) that hasn't bought the farm *yet*...Not sure if it's because it's a mobile chip with its higher heat rating on a PCIe card and huge heatsink. All my other GPUs (Radeon, Nvidia) lasted like two years before they started crashing/artifacting.

The big heatsink is very likely the reason why.
Also... and this is purely my speculation... but I think ATI might have paid extra-special attention to the HD5770 chip, because IIRC, this is the GPU that is used in both the Xbox One and PlayStation 4. I'm sure Microsoft does not want another Xbox 360 failure scandal (not that the PS3 was much better with its nVidia 7800 GS -based GPU and NEC/Tokin hybrid caps, among other issues). So that might be the reason why your HD5770 may be a little bit more reliable than the rest of the pack.

Also, I think ATI hardware generally improved quite a bit in terms of reliability starting with the HD2k era and going up to the HD5k era. HD6k was when things seem to have gone back to shit again. On that note, if you get a low-end HD2k or 3k video card like the HD2400 Pro/XT or HD3450, don't expect it to fail any time soon, even with its stock cooler. These cards are going by the truckload on eBay, and seems that almost none have failed. I have box of these as well. In fact, I did quite a bit of gaming on a ATI HD2400 XT video card back in 2014 when I had a little bit more free time again in college - probably second most I did in terms of hours, after the Radeon 9200 SE mentioned above. Best of all, I didn't mod that HD2400 XT at all, and I pushed it *hard*. Never ran above 67C, even in the hottest summer days with temperatures inside around 30C. While I don't consider 67C to be a good result at all, it's still rather impressive for a stock cooler on a GPU under 100% load at those ambient temperatures and for 2+ hour gaming sessions. Most stock coolers, I find, can hardly keep a GPU under 70C at full load. With the older nVidia GeForce cards like the 8000, you'd be lucky if you get less than 80C under full load. The single-slot 8800GT and 9800GT cards were possibly some of the worst offenders, reaching up to 90C easily with a hard load. No wonder they failed so much. Same goes for the 8400 GS with their tiny undersized/passive coolers.

[aztekk]

Re: Average lifetime of your discrete video cards?

I've been buying 3D accelerators since the days of S3 Virge, had 30+ cards/GPUs and the only thing I've had that failed on me was a Geforce 8400M, which after 5 years of use began displaying artifacts on the screen and was unusable. But I've never had an actual AGP/PCI-E video card fail, even the high end ones like 6800 Ultra, 7800GTX, etc. survived and are still working fine. Cooling fans have worn out, but not the chips. I don't know what you people do with your cards, but I am quite careful with my GPUs and other hardware in general. If you keep them cool, don't overclock, use an fps limiter and/or undervolt and don't do 12 hour gaming sessions every day they will last as long as your CPUs do, no rocket science... (assuming it isn't faulty by design, like the Nvidia 8/9 series or has bad caps)

It's necessary to note however, that unlike a lot of people here I do not allow any used components on my main systems, only on retro/backup PC's because I have no idea how the previous owners treated them and I have enough random issues with new components already. These days I also play games (not counting Solitaire) on average less than 30 days a year, maybe this contributes to my good experience so far.

I would really hate if my R9 280 (bought new in 2014) failed though, as I'd have to settle for an old HD 3000 series or try to look for faster NOS cards that work with XP.

[dmill89]

Re: Average lifetime of your discrete video cards?

Hmmmm..... hard to give an answer to this poll. Title says "lifetime of your discrete video cards", but poll has an option "Only had integrated, never had failure". Seems like it's pulling in two different directions.

X2, I've had a couple of the infamous GeForce 8x00 series failures, and a few AGP cards fail (I think a Radeon 9700 and a GeForce 5600 Ultra but it's been awhile). Otherwise I haven't had any failures and I've had mid to high-end PCI-Express cards in my "main-rig" since 2007 (Radeon HD3850, Radeon HD5870, Geforce GTX570, Radeon R9 380, Radeon RX470, GeForce GTX1070), with a mix of low and mid-range cards before that (which I can't remember the exact model of all of them).

[eccerr0r]

Re: Average lifetime of your discrete video cards?

Oh sorry the "only had integrated" option is for those who only have integrated; if you have at least one discrete GPU, then you're automatically disqualified from that option

And yes, my RadeonHD 3650 died on me, so I was a bit hesitant on the RadeonHD 5770 (this was the 3650's replacement!!!); but so far, again knock on wood, it's still alive...

[petehall347]

Re: Average lifetime of your discrete video cards?

i find the fan fails first .

[rugger]

Re: Average lifetime of your discrete video cards?

Hmmm, that is a pretty difficult question.

I remember only a few cards failing on me, including:

a) Nvidia 7300LE - Bad capacitors (the good old surface mount electrolyic's that look like polymer caps)
b) ATI 5850 - Just stopped working after a few years in storage. No idea why.
c) ATI 5450 - Artifacts and other weird problems.
d) Nvidia 9600GT PCI-e - Probably a bumpgate failure.

The rest of the cards I have owned either still work or I've given away years ago as I outgrew their usefulness to me.

[RJARRRPCGP]

Re: Average lifetime of your discrete video cards?

a) Nvidia 7300LE - Bad capacitors (the good old surface mount electrolyic's that look like polymer caps)

GeForce 7-series are known for having shit caps! Even eVGA was caught using Sacon FZs!

[rugger]

Re: Average lifetime of your discrete video cards?

GeForce 7-series are known for having shit caps! Even eVGA was caught using Sacon FZs!

Yep, was looking through my memory for that capacitor. The 7300le had those Sacon FZ's all over it. Looks nice when new, but the caps were junk.

[RJARRRPCGP]

Re: Average lifetime of your discrete video cards?

Yep, was looking through my memory for that capacitor. The 7300le had those Sacon FZ's all over it. Looks nice when new, but the caps were junk.

At least I never recall seeing that kind o' bullshit with the GeForce 9500 GT. (And I didn't with other 9-series from eVGA) (That probably leaves the eVGA 8400 and 8600 as the only ones to maybe still have FZs. If they even exist.....)
I had an eVGA GeForce 9500 GT, (2008) which was slow, (compared to even the GeForce 9600 GT and many later lower-budget video cards) but reliable... (It seemed that at that point, only other brands used FZs.)

[momaka]

Re: Average lifetime of your discrete video cards?

I've been buying 3D accelerators since the days of S3 Virge, had 30+ cards/GPUs and the only thing I've had that failed on me was a Geforce 8400M, which after 5 years of use began displaying artifacts on the screen and was unusable. But I've never had an actual AGP/PCI-E video card fail, even the high end ones like 6800 Ultra, 7800GTX, etc. survived and are still working fine. Cooling fans have worn out, but not the chips. I don't know what you people do with your cards, but I am quite careful with my GPUs and other hardware in general. If you keep them cool, don't overclock, use an fps limiter and/or undervolt and don't do 12 hour gaming sessions every day they will last as long as your CPUs do, no rocket science...

Well, the last few sentences there may be an answer as to why your video cards lasted longer than other people's.

Under-volting a video card (or any chip for that matter) like you suggested above, is known to reduce the thermal output (sometimes dramatically) of the chip. This can result in much much lower temperatures under load. But under-volting is technically modifying the card (be it through hardware or software). I bet you wouldn't get the same lifetime if you left the card stock / as-is.

Likewise, using a "frame-limiter" or simply using v-sync when the video card is capable of running at a much higher FPS than your monitor's vertical refresh rate will also make the GPU "idle" between frames, which will also result in much lower GPU usage and thus lower thermal output. For those of you who are curious about this, pop-out your favorite version of GPU-Z (but preferably a newer one that can show you GPU usage), start an older game that your video card can handle quite well (i.e. at framerates 1.5-3x higher than your monitor's refresh rate), and look at the GPU usage with and without v-sync. For more accurate results, turn off Anti-Aliasing (AA), as this can make some (older) video cards report 100% GPU usage when that's not the case. If you have a Kill-a-watt meter, you should be able to see a fairly proportional difference in the power draw too.

Also, while I don't like to assume anything... the fact that you live in the state of Minnesota suggests to me that your average room temperatures are quite possibly lower that other people's (especially compared to people living in the tropics or the "deep south"). This alone can affect the GPU temperatures enough to make a difference of a few years or more in the lifetime of a video card.

In particular, I've found out through experimentation (I've been "playing" quite a bit with video card cooling for the last 3-4 years), that room temperatures directly affect the load temperatures of a video card, especially on "dumb" coolers that run at the same speed all the time or multiple speeds depending on the load/2D-3D mode rather than the temperatures (traditionally the older gen GPUs) - for these video cards, the load temperature variation is proportional to the rise/fall in room temperature.

But for newer ones that vary the fan speed according to the temperature, the maximum load temperature will depend on the pre-programed cooling profile... and with that, some GPUs will run at somewhat decent temperatures and others just a hair below their maximum allowable temperature (think single-slot cooler GeForce 8800 GT and 9800 GT here, as well as 8800 GTS/GTX/Ultra and single-slot cooler Radeon HD4850, just to name a few - hence the failure rate on those).

So to help avoid anecdotal reporting in this thread, I think it would help if people posted here what temperatures their GPUs typically run at when idle and at 100% load for 1+ hours. Not only that, but posting your average ambient/room temperatures, as well as posting your video card's information and details about what kind of cooler it has will certainly give a clue as to why it might have failed or is still working.

On a finishing note, here's something to think about: how come (desktop) CPUs fail so rarely, but GPUs fail so often? Does the load temperatures have anything to do with this? Or perhaps the socket? (The fact that most desktop CPUs don't have to go through a harsh heat-cycle for BGA soldering???)