AMD Cards Don't Last As Long as Nvidia Cards. AMD Cards Die From Overheating.

[Batou1986]

Strange the only GPU's I've had to replace because of failure in the last 10 years where made by Nvidia and that's out of about 20 cards most of which where AMD/ATI

 

[jboydgolfer]

Someone better tell my 6950 that is in a daily runner, to go out and get a TINY fire extinguisher.... Oh, it's being used by my kid to run GTA5 ,and its running it quite nice too, for hours, and hours.

But, to each his own. Im no fanboy, if it works im happy, this just happens to be the situation in MY case.

 

[flmatter]

I've heard that AMD cards overheat a lot and die more than nvidia before, so it was interesting to hear someone that deals with an insane amount of graphics cards confirm it.

It is on the internet and youtube, so it is confirmed and true by one person that does nothing for the industry. Trolls were fed

if it is on the internet it must be true

 

[silentbogo]

Regarding this topic: not exactly true. This guy only messes with High-end cards and in my personal experience - NVidia has significantly higher failure rates in low-to-mid price segment. One of my broken hardware suppliers just sent me a list of 100 cards that have been returned by customers and top of the list - NVidia GT630, GT730 and 740. The main cause ranges from typical GPU failure and VRAM failure to a complete burnout for reasons now unknown.

Last month I had a similar list with low-end NVidia and high-end AMD at the top (R280X and 79xx mostly).

Also it is worth mentioning that ~80% of observed failed cards are manufactured by Powercolor, HIS or Zotac - you rarely see a reference NVidia or EVGA cards failing within first 5-6 years. In a few weeks I'm gonna get another list with tons of MSI cards, we'll see what's happening there. And even though I have some expertise in this area, I can only assume/suggest that the cause is mainly bad cooling or bad design. It might as well be "solar storms" or "global conspiracy".

BTW I still have a EVGA GTX275 running in my stepdad's rig, while a pair of HD4850s powers my brother's PC. Two HD3870s and a single HD3850 are collecting dust in my closet, but are still fully working and very much OC'able.

 

[Nosada]

I like how some people seem to think that because they have used "100's of cards" their findings somehow stop being anecdotal.

I have not seen a single source that places failure rates of GPU's anywhere but real close together. The only recent memory of mass GPU death I have is when someone forgot to enable the thermal protection in a mobile driver, and guess what, that was nVidia, not AMD/ATI.

Here's my anecdotal "evidence": in the several hundred PC's I've built for family, friends and acquaintances, none has ever died without being tampered with (physically or driverwise). That obviously proves video cards NEVER break, are bulletproof, and can kill Superman AND Batman at the same time just by spinning their fans.

Now let me go check on my backup server sporting a X800GTO2 flashed to 850XT 10 years ago ... oh look, it's still running, how quaint, it must not have gotten the memo that it should already be dead.

 

[HumanSmoke]

Also it is worth mentioning that ~80% of observed failed cards are manufactured by Powercolor, HIS or Zotac - you rarely see a reference NVidia or EVGA cards failing within first 5-6 years. In a few weeks I'm gonna get another list with tons of MSI cards, we'll see what's happening there. And even though I have some expertise in this area, I can only assume/suggest that the cause is mainly bad cooling or bad design.

Bad design usually covers it. Hardware France have a six-monthly roundup of returns to a large European components supplier, and the list of affected graphics cards (which would include other issues such as noise, general customer dissatisfaction, coil whine etc.) tends to be weighted towards cards that either have a reduced bill of materials, or (over)clocks set too high for stable operation (as is the case with the Gigabyte GTX 780 Ti Ghz Edition in the latest article).

 

[R-T-B]

People have seen most dead AMD cards because of the bitcoin mining craze where pretty much only AMD cards were used. Of course such cards die, they were never designed to grind at 100% load 24/7...

This. They are buying used cards that were certainly used for stressful bitcoin or other crypto mining (nearly 100% of that gen were) and then complaining that they are broken.

Also, Powercolor, Zotac and HIS were goto mining brands because they were cheap and had decent ram setups that did not prevent mining.

-RTB, former GPU litecoin miner, and ASIC bitcoin miner

 

[qubit]

People have seen most dead AMD cards because of the bitcoin mining craze where pretty much only AMD cards were used. Of course such cards die, they were never designed to grind at 100% load 24/7...

It would be interesting to see what improvements would need to be made to build quality for reliable 24/7 100% use. Oh and how much that would cost.

This applies to both AMD and NVIDIA.

 

[RejZoR]

100% constant fan speed and fans with HDB bearings. Or a massive cooler as found on Strix/Gaming/WindForce/TwinFrozr cards. But these were more expensive aftermarket ones which were not feasible for cheap mining. Stock reference coolers usually ran on temperature limit and the fans often crapped out because of it. I do have to admit that reference Radeon blower fans were usually really poor quality and they started rattling quickly.

 

[natr0n]

Any card will die if you game hard over long periods of time.

FYI...I've killed two 7870 hawks.(vram went bad on both.)

 

[Mussels]

Though from the sounds of it he can't find these cards anywhere.

That is interesting.

AMD has an 18% market share. he should find less than 1/5th as many cards for sale as he does Nvidias. Anything else is rather idiotic speculation.

 

[TheoneandonlyMrK]

It would be interesting to see what improvements would need to be made to build quality for reliable 24/7 100% use. Oh and how much that would cost.

This applies to both AMD and NVIDIA.

As a folder ive ran both brands for years 24/7 and by looking after them none have ever died, purely a click bait exercise this.

 

[john_]

The only cards that we know to die after a period of time, I mean it is a fact that they will eventually die, are Nvidia 8000 and 9000 series cards.

 

[silentbogo]

It would be interesting to see what improvements would need to be made to build quality for reliable 24/7 100% use. Oh and how much that would cost.

This applies to both AMD and NVIDIA.

Probably nothing. I'm pretty sure that enthusiasts would still use expensive aftermarket cooling and occasionally find ways to kill GPUs, while mortals would still use them as is and occasionally find creative and funny ways to kill their GPUs.
And sometimes you'll see GPUs die of their own accord. It's just how things work.

AMD has an 18% market share. he should find less than 1/5th as many cards for sale as he does Nvidias. Anything else is rather idiotic speculation.

He's either lazy or stupid. Retailers have no motivation in keeping the entire lineup of old graphics cards(even refurbished), so I don't know why jayz mentioned it in his review.
Betcha I can even get a slightly used ATI Rage 128.... oh wait... I actually do have a slightly used Rage 128 somewhere in the box with all my other ancient crap.

 

[FordGT90Concept]

I've bought almost exclusively ATI/AMD GPUs and none have failed outside of warranty. I retired a single 9800 Pro simply because the fan on it died and didn't bother replacing it. It was still fully functional, just really hot and AGP.

I still use an original 32 MB PCI Radeon on occasion.

 

[tabascosauz]

What is trying to be proven here? That AMD silicon is somehow inferior under high temperatures even though they also come from TSMC?

Have we all forgotten what a shit-show the GTX 480 was? 96 degrees? It would be completely unacceptable to leave the 480 out of this. AFAIK the 290X was the one card that, under a reference cooler, could even come close to that kind of scorching temperature.

It's like pointing to a 4790K that was force-fed 1.4V+ and blaming Intel for making a weak CPU that insta-died under the voltage, while pointing to a 4770K that was able to run 1.4V for about 2 days.

 

[gaximodo]

I don't think most of us expect a video card to last 4 or 5 years. We move quite easily nowadays

I'm sorry I might be greedy but I expect my high end graphics cards last ~10 years. They can be not keeping up with current titles in terms of performance but this is not a reason why it should die within 3 - 5 years, especially other components of the similar quality lasts ~10 years.

I don't think this thread will be any useful other than a reminder that this is another aspect for ppl to consider while making purchase decisions.

I think it's safe to say nowadays AMD tends to push their GPUs' harder than its competitor, which results a smaller overclock headroom, and potentially worse life expectancy of their GPU's.

 

[dorsetknob]

And just got one thing to mention Solder problems now which brand had had this problem BAD in the late 90s/ early 00s chips and chipsets just died untill people found out a reflow would repair the card
lets start with the 7 series and then the 8 series Nvidia cards
The op probably to young ( mentaly )to Remember the fuss this caused
ps Repair shops are still dealing with the fall out of Soldergate
Now how many Millions of $ did Nvida set aside to sort that out
and how much did it cost AMD ?????????????????????

edit never destroyed an ATI or AMD card or had one self destruct on me
been useing ATi/AMD since the 3dfx days
my cards dont die they get upgraded and then sit on a shelf for nostalga (fully working ).

Never dumpster dived a Bad AMD Card Either got pleanty of Nvidia bad one tho

 

[64K]

What is trying to be proven here? That AMD silicon is somehow inferior under high temperatures even though they also come from TSMC?

My guess is this thread is a Nvidia fan payback for the posts about some Nvidia drivers were killing GPUs on the AMD driver fix thread in the News section recently. The kind of threads like this one usually turn into a shit-fest pretty quickly but who knows.

I'm sorry I might be greedy but I expect my high end graphics cards last ~10 years. They can be not keeping up with current titles in terms of performance but this is not a reason why it should die within 3 - 5 years, especially other components of the similar quality lasts ~10 years.

I'm sure that cards could be built using the kind of quality parts that would ensure that most would survive for 10 years of use but most people upgrade their card long, long before that. Would you want to pay extra for the kind of parts necessary to last 10 years for every card that you buy knowing full well you won't be using it for 10 years anyway? I doubt most customers would and hence it's why it's not a priority of either Nvidia or AMD or their board partners.

 

[jboydgolfer]

If there WERE any actual solid tangible proof of this , I would Certainly cite that AMD users (atleast advanced ones) are more likely to flash they're gpu's to try and get a little more HP out of them, as with the 6950/6970 Bios mod's of 2010 on.just seemed relevant to point that out to me, and it couldn't hurt to mention the fact that amd gpu's are commonly the choice of small scale Bitcoin miners too, which if you know even a little about the process, causes a Great deal of wear and tear. Im sure one has more than the other, but I dont know if its SO cut and dry that You could just put a label on it like "AMD gpu's are More likely to break because.......... reason"

 

[jjnissanpatfan]

I do not believe this to be true. I have always monitored temps and what not, so my cards always last. I know my 800xl works, and my friend still has a 9800xt. No matter what card you have if let dust build up and do things you shouldn't you will have problems.

 

[hat]

Everyone puts out shit products (in this case, cards that overheat). AMD is/has done it, and Nvidia is guilty of it too. Just because Maxwell runs cooler than a controlled nuclear reaction doesn't mean they've always made ice-cold cards, and that doesn't mean AMD cards always catch fire as soon as you take them out of the box.

 

[dorsetknob]

 

[Solaris17]

I didnt even watch the vid the name is too ridiculous. Does he pass himself as a professional reviewer or something? Either way oddly enough Iv only had nvidia card die, Actually thats not true my ratio is about 4/1 in my GPU history, I nuked a 2600XT once because the fan failed on it. Which honestly im not sure counts as the card itself? /shrug

At the end of the day im just going to need something a bit more substantial than some youtube guy buying used GPUs named jayz2cents or some shit. For now I'll go with personal experience.

 

[rtwjunkie]

This guy is nuts!! How often do we hear about ANY cards dying or overheating from EITHER brand on forums, including this one?

I'll answer that myself: Not Very Often. When they do, it's almost always the user's fault, either by failed vbios flashing, overclocking/overvolting, or running in cases that have terrible airflow.

I believe 90% of the overheating failures could be averted with setting up fan profiles on their GPU's and getting better cases.

Although about 80% of my cards have been Nvidia, I have to say, the only GPU I've ever had go belly up because it just sucked was a Zotac 9600 GT. That's an Nvidia, for those that don't know.

This troll thread need to be closed.