• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Nvidia nerfing mining on their GPUs

Since we're trading anecdotes here, plenty of us on this very forum have bought ex-mining cards which have died soon after. Myself included.

The memory on mine died after a few weeks of having it. Quelle surprise. Won't be making that mistake again.

Anything I would post in response to that is already in the post you quoted.

I don't buy 2nd hand cards for myself either way, and I don't sell mining cards.
 
I have multiple cards yes, enough that "mining reduces the lifespan is a FACT" I should have had at least one failure, I've had zero.

In the same time I've had 1 out of three gaming only cards fail.

If you set them up correctly, and know anything at all about electrical engineering, it should be obvious that cards running at 120% of their rated power with constant spiking and changing of fan curve, are exponentially more likely to fail than cards being run at 60% with a fixed fan profile.

Hyperbolic references to cancer are meaningless.
I don't mean to dig but I feel like a balanced viewpoint is essential.

You're talking to a guy who gained an Electronic Engineering degree before going on to spend over three years working for MSI. I've also done a bit of mining myself in the past and spent a few days with guys operating a mining farm of thousands of units. I understand the issues a lot better than you seem to think. You, however, either don't or, more likely, you are attempting to mislead people when you say that you run the cards at 60%.

Yes, most miners run the GPU underclocked and undervolted. Yes, absolutely, gaming is a lot more stressful for the GPU than mining but it's not the GPU that's the issue with mining, is it?

What you have is a graphics card designed for gaming, made by a manufacturer who is expecting the GPU to run maxed out for most of its life, dealing with random spikes, peaks and troughs of load, but expecting the memory to be at less than 80% most of the time and operating relatively sedately in comparison. Said manufacturer doesn't want unnecessary cost, and as a result, they usually have a hell of a lot more, & sometimes, better quality, VRMs providing power for the GPU than the Memory and always, much better cooling on the GPU than the VRAM.

What mining does is flip that expected use case on its head. As a result, running heavy load through the memory and it's lonely VRMs. Mining puts more heat and more load through components which weren't designed for 24/7 maximum performance and that's where the failures can come from. Obviously, how well ventilated the cards are will have an impact on that failure rate since silicon hot spots are lethal, but it doesn't completely remove the fact that the card is operating outside of the manufacturers expected design parameters when it is being mined on. To me it's kinda like buying a car and driving it everywhere in reverse.

The very simple fact is that most graphics card failures, whether they failed through gaming or mining, fail because of the memory, not the GPU. Mining puts extra stress on the weakest parts of the card's design.
 
Last edited:
I don't mean to dig but I feel like a balanced viewpoint is essential.

You're talking to a guy who gained an Electronic Engineering degree before going on to spend over three years working for MSI. I've also done a bit of mining myself in the past and spent a few days with guys operating a mining farm of thousands of units. I understand the issues a lot better than you seem to think. You, however, either don't or, more likely, you are attempting to mislead people when you say that you run the cards at 60%.

Yes, most miners run the GPU underclocked and undervolted. Yes, absolutely, gaming is a lot more stressful for the GPU than mining but it's not the GPU that's the issue with mining, is it?

What you have is a graphics card designed for gaming, made by a manufacturer who is expecting the GPU to run maxed out for most of its life, dealing with random spikes, peaks and troughs of load, but expecting the memory to be at less than 80% most of the time and operating relatively sedately in comparison. Said manufacturer doesn't want unnecessary cost, and as a result, they usually have a hell of a lot more, & sometimes, better quality, VRMs providing power for the GPU than the Memory and always, much better cooling on the GPU than the VRAM.

What mining does is flip that expected use case on its head. As a result, running heavy load through the memory and it's lonely VRMs. Mining puts more heat and more load through components which weren't designed for 24/7 maximum performance and that's where the failures can come from. Obviously, how well ventilated the cards are will have an impact on that failure rate since silicon hot spots are lethal, but it doesn't completely remove the fact that the card is operating outside of the manufacturers expected design parameters when it is being mined on. To me it's kinda like buying a car and driving it everywhere in reverse.

The very simple fact is that most graphics card failures, whether they failed through gaming or mining, fail because of the memory, not the GPU. Mining puts extra stress on the weakest parts of the card's design.

My cards, how I have them set up, the memory and VRM's are running cooler than my gaming cards do at full chat, it makes less than 1MH difference to my mining scores to have them running cooler, which is literally pennies a day I'm losing. Yes I do know what kills GPU's, and it isn't mining, its mining (or gaming) with no regard to the lifespan of the card, but software is just software, setting up your hardware to run that software is the same either way. Running software doesn't kill graphics cards, people with bad setups kill graphics cards, but that can happen regardless of what software they are running.

Gamers overclock their memory too, and as you say many cards have very poor or even NO cooling on their memory. Thats not a design feature, its just bad design, so if memory is failing on cards because of poor cooling I'm not 100% sure why "miners" are being blamed for this, because as you say, gaming is doing this too. Mining is just CUDA software, and yes cards are sold as being capable of CUDA, so it is a bit of a cop out by the manufacturers to say "oh yeah, we didn't mean THAT CUDA".

To be clear, I'm not suggesting people should rush out and buy 2nd hand mining cards, no more than I'm suggesting people should buy any card 2nd hand with no warranty, I was only responding to someone's comment that mining specifically reduces the lifespan of the card and gave 5 years as an appropriate lifespan for a GPU, saying that "mining" kills cards in 6 months.
 
Last edited:
Wonder how many gamers are running their cards 24/7 for gaming... that's a lot of Mountain Dew :p

But I know what you'll say, then it becomes about "power cycling" and "heat cycling" being worse than constant 24/7 usage... we've heard all the arguments before.

e: The simple fact is, you'd be much better off buying a 2nd hand GPU from a casual/intermittent gamer than a hardcore 24/7 miner. One will be in a lot better condition than the other.
 
Wonder how many gamers are running their cards 24/7 for gaming... that's a lot of Mountain Dew :p

But I know what you'll say, then it becomes about "power cycling" and "heat cycling" being worse than constant 24/7 usage... we've heard all the arguments before.

e: The simple fact is, you'd be much better off buying a 2nd hand GPU from a casual/intermittent gamer than a hardcore 24/7 miner. One will be in a lot better condition than the other.

I already said repeatedly, I'm not advising anyone to buy 2nd hand GPU's, I responded specifically to someone saying that "mining reduces the lifespan of a GPU is FACT". And that miners GPU's fail in 6 months and that miners put in 10 Rma's in 5 years.

The fact is, you are never going to know how a card has been treated until you buy it and it fails on you, so how about just not buying 2nd hand GPU's and spend the equivalent money with a retailer so you get like 3-5 years warranty instead?
Then you'll say "scalpers are buying all the cards", yeah same thing, card could still be DOA and if you bought it from a scalper you don't get warranty either.
 
Last edited:
The wear on the gaming gpu also depends on the games the person plays and if they have upgraded the gpu after a short time of using it and had it in storage, eg the Vega 56 is generally said to be power hungry but for most of the games I play it runs at well under 100watts, especially with Chill on. Also I upgraded a 680 within 2 years, kept it in storage as a back up, sold it last month. So although it was a 9 year old card it had only been used for about 2.
 
No it is just supposed to reduce mining, nothing else. Even then I have heard that it might not impact the mining of most coins, but perhaps Nvidia could update the drivers to reduce mining on other coins if they wanted. Nvidia announces the GeForce RTX 30 series LHR graphics cards - Graphics - News - HEXUS.net

If they include more coins in the drivers, miners will just use the older drivers, so it wouldn't really gain anyone anything - they would have to release new hardware to include new limiters so that working drivers were not available
 
Back
Top Bottom