Crypto is unlikely to crash meaningfully anytime soon in a way that would make everyone sell their hardware, and if anything, a crash encourages just as many to mine more as difficulty may drop + being in it for the long term, but ethereum switching to proof of stake is going to move a lot of GPUs around from miners who aren't interested in mining lesser coins.
@@boiiifuaintcare8629 lol tell them 🤣 laughing at all those who think it's gonna flop crypto has proven at this point if it falls it won't crash as hard as we have seen in the past. So while the future of mining isn't as clear the future of crypto doesn't seem too bleak
Yup, when GPU prices crash == when Crypto crashes, it will happen every couple of years, if you miss your shot then, you need to probably wait atleast another 3-4 years.
Bought my 3070 from a miner, 1 year used card undervolted, mining rigs kept in a 15c temp enviroment. The person even showed me his rigs and how he takes care of them. I also got the warranty papers and i couldn't be happier with the purchase. Conclusion: If you know that the person takes care of the cards, i'd rather buy used from a miner than a gamer.
that's 100% what I heard everytime when I did proper research. there's a myth of "oh my God he's using the board 24 7" yeah and the board is undervolted and most times temp checked.. we're running gpus for 10h a day running at 100% and 80º 😂😂 which is most destructive? really like some people like to hear buzzwords and jump to conclusions instead of thinking
As a gamer, i can say that i didnt take good care of my rig cuz im not using it to make money So the miner's card is most likely in a better condition than mine...
@@bbblop4545 well it doesnt run 100% 24/7 cuz I only play for maybe 2 - 3 hrs a day at max Just dont have that much energy to play after work But I never really clean it and sometimes it makes weird noises Im not sure how the memory gets cooked, am not familiar with how the hardware works :\
The most likely "degredation" is simply the lifespan of the fans themselves. If a fan motor is rated for 40,000 hours, that's about 4.5 years of 24/7 operation, so if you buy a used miner's card that's been going 24/7 for a year, those fans should be good for another 3 years of continuous 24/7 use.
Should be noted that the motors are rated for normal usage cycles, ie more stop-start than 24/7. They can actually last longer by virtue of never stopping. But iirc this effect is only in the single digits percentage, so it probably gives you a week or two of fan time, hardly noticeable. And then if they wear out just ziptie some noctuas on and you're good.
@@TheRacerperson I said it was one of the reasons, the other was the rest of the team started to suck, grinding for tanks toolk away way to much of my free time. I owned several T10 tanks. I stopped when they added the Japanese tanks, which is several years ago.
WOT is and for a very long time has been pay to progress, not pay to win. Dosn't matter how much money you throw at the game, if you suck you'll still suck.
Palit used to be "Avoid on sight" until the 10th generation - thats when they stopped catching fires from nowhere and actually got some nice thoughtfull heatsinks and quality components.
miners are the reason why there are gpu shortage and why pricing is what is they are right now. just don't buy anything from them. let them soak in the cost as the result of their selfishness.
@@vaynelevant Have you not noticed that there's, I don't know, a pandemic going on that's causing a semiconductor shortage? If the problem was miners, why are all computer parts having a price increase or, for that matter, why are _vehicles_ having a price increase? Demand is high for these products, but supply is crippled. Nvidia and AMD know after 2018 how much demand mining creates, so they adjust their foundry booking accordingly. The problem is foundries downsized during the pandemic, and are having trouble scaling back up. Moreover, miners wouldn't buy cards at high prices. They have to make money after all. Would _you_ buy a card for $1000 that would normally be $500 if you were trying to make money with it?
@@coolmemesbudd There is a reason that i can buy a 3080 at a retailer now (even though for inflated prices), but couldn't 4 months ago. That reason is, that Ethereum, the best cryptocurrency to mine atm, will be switching to proof-of-stake soon. High investments in new GPUs will not pay off for smaller businesses anymore. The biggest reason for inflated prices definitively are miners, with people buying from scalpers on the second place.
If Palit is testing and claiming this with their cards, I will assume it's a problem with no other brand and will avoid theirs altogether. Good information Palit.
They could just be referring to cards that have their thermals degrated due to heat which results in the card throttling sooner. Usually repasting and re-padding them will restore performance. DIY Miner/Gamers dont do this often but professional miners take care of their assets.
My theory is that crypto mining despite using the gpu 100% 24/7 doesn’t usually start and stop the load like normal games, so the card doesn’t change temperatures too much. Less temperature change means less thermal expansion cycling and I believe this is why mining cards probably don’t degrade as much as people would think.
its the same reason old taxi's and stuff can reach 500,000 miles without any major issues becuase the motors are hardly ever stoppped and started they just run all day long arond a city
As a proud owner of a 2080 from Palit: Card still runs smooth, unlike one of its fans. Nothing some percussive maintenance couldn't fix, but a few months ago (when GPU prices where peaking) it gave me a real good scare for a good hour.
Ethereum mining will stop when eth2.0 releases, and when the crypto bear market comes. Of course a miner with 10+ cards will sell them if they are not profitable anymore.
It is inevitably coming, the big coin that most of these gpus are mining is about to go proof of stake within the next couple of years and that will undercut the value of gpus for crypto even if the market never crashes.
A smart miner usually underclocks his graphic card, so it uses less voltage, but still give a solid amount of mining power, so the profit actually is higher than it would be when using in stock settings. So a card that has been used heavily in current gen games, is most likely a worse option to buy in second hand, than a mining card, which ran underclocked in constant usage.
@@ssllsg9439 from room temp to max temp and then back to room temp is thermal cycle, when gaming your card went from 20celcius to 90 in game and then back to 20 when the game is loading-that 1 cycle, and like Mbdk2 said, it could be over 100 cycles per month for an average gamer
Another would be that if you all start buying up second-hand cards you will send a signal as a market that you're content with being second-hand consumers.
@@tomcat6186 You can work for money and game in your free time. Because resource and energy are limited, and we do not have a surplus of renewable energy still, mining is a form of speculation which is against common welfare. A Chinese study says that China cannot reach their climate protection goals if cryptomining is to continue.
Honestly, if they are selling them for 10-20% of the original (not inflated) price, through ebay where I can get refunded if it doesn't work, then sure, I will take a 2080 for $50.
Linus: "To finally turn the tables on those filthy filthy miners" Mom: "Sounds like he can't get his kids to take a bath either" Me: "Miners not Minors" Mom: "What??"
Gamers card: Full of dust, ran hot and turned on and off allot (turning on and off is the main cause of component failure) Miners card: kept cool, cleaned regularly and undervolted
@@whwhwhhwhhhwhdldkjdsnsjsks6544 dumb miners do that probably running them at 100% costs a lot more energy and can kill the cards so they have to reinvest xd
@@whwhwhhwhhhwhdldkjdsnsjsks6544 And in general second hand stuff is often damaged or in bad shape. What miners and people in general are doing with unstable cards etc.? They are selling them! Another thing is that people are planning this from the start. I saw how render farms do it - they don't care if processors or graphic cards survive for long time because they will have new faster stuff soon anyway and components that have some issues are slowly sold one by one. There are professional scammers who are selling whole PC builds with swapped components. Majority of people know nothing about computers and it's insane how often it's exploited. You can find shell companies that use "hit and run strategy" - they are selling horrible shit on mass, then close bussines and open new bussines again...
honestly at this point, I dont care if old mining cards preform just as well as a new card straight out of the box, Im not buying a used mining card. Miners and scalpers have screwed the graphics card market for over two years now to where I havent been able to upgrade from my old 970. Im not buying something these guys are selling second hand. Period.
I think miners and scalpers, while exacerbate the problem, are not necessarily the sole responsible for screwing the market. Theyre just the second in line. Global shortage of silicon, manufacturers not able to make supply, high demand, rapid technological growth are the main ones. Im saying this as a reminder to others and to myself mostly, its easy to forget
@@WingMaster562 yeah it's usually multiple unexpected problems when a business fails that hard to meet supply. They probably had plans for the crypto bubble because it wasn't completely unprecedented
They run hot and at high heat so it’s not a performance issue but longevity. Heat kills components faster so the card is way more likely to fail sooner, especially since the last wave of miners are outdated fad idiots who somehow convinced themselves inefficient GPU mining gets more bitcoins than SHA256 ASIC rigs.
I never expect the GPU core to be the part that dies in old mining cards, I would be more afraid of the power delivery giving up after running 24/7 for years in possibly hot temperatures. Or the memory dying. Or the fans dying (sure, they can be replaced but it's quite a hassle to find replacement fans and it costs extra). It has to be REALLY cheap to be worth the risk as Linus said. I'm not gonna be accepting a 10-20% discount for the unknown reduction in lifespan of the card.
Mining is almost as intense as gaming but it's done 24/7. As long the thermals for the card are good, they should be ok. Most miners wouldn't fry their card cuz then they'd have to buy new cards and lose profit
@@ZastropollyonZ "Most" is not good enough for investing in hundreds worth of dollars. What if that "most" fails? It's a gambling at this point, a bad one at that.
I wish i did when the price was right. Instead i bought a 1070 for only $100 less than the 1080 ti price months before. Even then my 1070 was way cheaper than the horrible prices of today. So glad i didn’t hold off replacing my r9 270. Otherwise i would have played all my current games at low to mid settings.
what about the memory frequency of mining GPUs?does it reach the advertised frequency?, the most component that dies on these GPUs are the memory chips, for example the rx 580 which is used a lot by miners, has no vram sensor, the gpu core's temperature could be in the safe range, but vram will be extremely hot because mining stresses the vram more than the gpu core
If the gpu memory wasnt in the safe range it would throw errors or perform slowly and the miner would be taking a look at the card. Typically only memory which is bare with no fan will be at risk of degredation or failure.
on most of the 30 series card i’ve seen and held, the ram is directly onto the plate of the heatsink just like the gpu die, so generally you get the same cooling on the memory, and since the die is sipping power, the most heat producing thing becomes the memory, which is cooled by the heatsink. so i’d thing they’re fine. plus most programs show the vram temp and have alarms.
I have had 3 rx480s that I bought from miners that have failed due to vram artifacting, its plain and simple, wait it out, dont buy used cards from miners
Mining puts the main pressure on the RAMs while the fan rotation speed and card temperature are based on the GPU temperature. Because the GPU does not work much, it causes the temperature to be low and the fans to operate at low speeds, while the RAMs receive very high temperatures due to heavy mathematical activity and calculations. They become very vulnerable and their useful life is greatly reduced and they become like coal
@@phenomenologicalparadox5216 Why? Because how they behaved and and laughed at people when the cards were expensive like never before and crypto didn't crash yet. I know that some of you have absolutely none dignity or pride in yourself, but i am simply not going to give them money, making them even more profits.
1. Mining GPUs are typically undervolted 2. They are on 24/7 for the most part and don't typically thermal cycle. As long as they're cared for and aren't dusty as heck and moderately taken care of it's fine.
Yeah... Very odd linus didn't mention that. The memory and memory controllers are the things that get cooked on mining cards. It's not a bad idea to look at the pcb and check for discoloration/signs of high heat around the edges of the die (the rubbery stuff), the memory chips themselves and the VRMs if the seller lets you. Avoided more than 1 bad card that way. It's an easy fix if the card still works, though. Lower memory clockspeed until the artifacts go away.
But it wasn't because it's mining I have one with artifacts and I've never mined in it, and I have a 1060 6GB that mines since 2016 and it has absolutely no problems, so much so that in some games I even use overcloks without any problem.
@@VEVOsdead True, but he neglected the most important word in those sentences, "memory" and then proceeded to solely focus on gpu clockspeeds which certainly can lead to artifacting, but no miner is pushing gpu clocks. The gpu pipeline barely, and I mean barely gets used while mining. That's the frustrating thing about LTT. They usually get the info right, but not how or why it's important in the first place aside from a ultra surface level stuff such as more clocks = more better which if you're watching this channel I would think you at least have some sort of understanding of. I mean I understand that's the target audience of the channel, but it frustrates the hell out of my how insistent they are with the "mainstream-ness". I'll still watch though. Also, furmark is a bad example because modern gpus will detect furmark and automatically lower clocks somewhat. Trust me I've done everything Linus did here and it's not a reliable way to tell if a gpu is "good". Had gpus pass as much furmark as you can throw at them only to crash the instant a 3d frame gets rendered in battlefield. Furmark is primarily used to measure the absolute max power a gpu can draw. It's a power delivery test, not clock stability test. A 30 min loop of timespy extreme is way way more predictive if you actually want to test the clock stability of a card.
Good point about the maintenance. I would suspect a lot of the miner cards are in a server room type environment and maintained well, so even though it's been heavily used, it's not caked in dust like a lot of home system graphics cards are. Thanks for the video.
See, I honestly appreciate this kind of video. Its information that many need and want but very few have access to, BUT at the same, time, the minute this video gains any kind of traction, 90% of the used card list will bump up in price, so it'll in effect not "fix" the lacking card problem *as much*. Its a double edged sword if i've ever seen one.
To be totally fair, in the grand scheme of things, it doesn't matter. Not having this info could do more harm than good. This info doesn't matter until the crash happens and, when it does, I would think people trying to liquidate cards, if it happened as fast as it did last time, wouldn't dare try to raise prices because their competitors might saturate the market before they do, and thus would be left with hundreds of cards they don't want. So sure, some people will bump up their prices with the trust in used cards going up, but until GPUs are readily available again GPUs weren't going to be affordable anyway. Any rise in price this video makes will be circumstantial at best.
you can have a severe degradation on performance from a card that has been running 24/7 for years, that's not a myth. it's called dust and dry thermal paste, just clean the fans and replace the paste.
Its a myth that the degradation is unrepairable. Not to mention hashrates are highest when Undervolted which means mining is easier on a GPU than gaming
Bought a rx570 for 90usd in last crypto crash. Still runs everything you throw at it med to high 1080 60fps Waiting for next crash will get a 3070 or something. 😎
I bought a used rx570 and it crashed the following week. I changed it with another rx570 that was used it also crashed. They wese both mined ik damn sure😂😭😭
Now you should do a comparison of a card that’s been used for gaming for three years, versus one that’s brand new. Guarantee that .8% is gonna look like a speck
@Алексей Незнамов yes, it can Happen, but it is extremely unlikely to happen. I had a friend that was addicted for a certain MMO, and he played it for about 18 hours a day, and kept this routine for ~ a year, then he committed suicide after the Company banned his acc
@@workout3D it doesn’t matter if it was used 24/7. Gaming causes massive transient spikes that mining just doesn’t cause. Even if you’re only a casual gamer playing on weekends, you can expect more silicon degradation over three years compared to a miner using it under relatively low load and undervolted for three years…
Mining etherium uses memory more than gpu power, so most mining cards would be optimized to run at a lower clock speed(and voltage) for the gpu and have the ram speed overclocked to as high as the card would support. It would be more likely for the ram to degrade over time than the gpu.
@@klyplays No, you don't need to be a pro at SMD. Also, ram tends to NOT die from an excessive overclock as much as it tends to just say "I'm not going to run at all, turn the clock down and I'm fine" since it normally doesn't run nearly as HOT as the actual GPU die.
I bought a second hand Ati radeon sapphire rx580 8gb Nitro+ that was clearly used for mining. It worked an absolute treat for just over 1 year before it didn't. At the time I needed the performance at bargain basement price, so looking back it was somewhat worth however I wouldn't ever do it again. Having that feeling that it could croak at anytime really sucks.
It's not the fact that it was used for mining that was the problem. it's how they took care of the card. If you keep thermals under 65 degrees the card will last for 5+ years easily. It could have even croaked from your own poor heat management. I have a R9 280x that mined for 4 years. I put a new cooler on it, changed thermal paste and gave it to a friend for his gaming PC. still works 8+ years later.
@@Bokuzen035 I agree however 1. Not everyone knows how to replace coolers and thermal paste especially in a gpu. 2. Mining cards are often left in opened cases and subject to environmental factors mine had rust all over it. Years of mining in a humid basement probably. 3. Wear and tear affects more than just a couple of components, I'm pretty sure that one of my transistor's was faulty when it finally broke down after a year. Each to their own, like I said at the time I needed something good and cheap but you pay for what you get. Personally I wouldn't recommend it nor would I do it again, save the few hundred and get a new one. I paid $150 for my rx580 8g when a year later I got a 1660ti super for $420 when it died. The reactions of owning a ex mining card is either "yay I got this beast super cheap" or "fuck when will it breakdown"
when the deflation occurs. But at the moment the demand is still high, the supply very low and too many people are willing to buy cards for unreasonable prices because they fear that the money will completely lose its worth over the next years.
@@vladpintilei6204 New silicon foundries won't be online until 2024, so it's possible it could last that long. Supposedly only 20% of the demand for GPU's have been met which means 80% demand still remains.
Never, but u can expect certain Categories of GPUs to appear. In reality 99% of consumers do not need 3090. For avg consumer 3060ti should be considered like super high-end and also super expensive. Often people just want to buy nice things they dont need like expensive cars. My housemate has 3080 and has 1080p monitor with shitty refresh rate, doesn't do any Work that requiers rendering, doesn't play newest games... What's the f-in point ? :D All of those items are Luxury items and while in great demand they are not needed ASAP.
i bought a 1060 6gb about 3 years ago, and before i bought it, it had been used as a mining card since new. Got a good deal, and it still works like new
They are fine as long as you supply them with decent PSU. Most GPU used for mining went underclocked to press its pressure, gamers and enthusiast on the otherside often overclock their GPU to achieve the maximal juice they were promised to deliver. The reason why miners sell their card cheap it's because they are on tight schedule on replacing their rig (time is money), plus they already reached 100% ROI from their card they are trying to sell. I commented based on my experience.
Counter argument: The miner friend has a strong incentive to provide the best case cards, so that they can continue to unload then on unsuspecting users. At least a new card comes with a warranty.
My main issue: Linus only tested the performance of the clock speeds. No tests performed on the CUDA cores, the onboard memory, or the fans. These are all things that may have aged as well, and independently of the core clock, at that.
Hey, previous full-time mining op manager with a specialization in immersion cooling chiming in here (this'll be a bit longer, sorry) : Adding to what others have said about fans being the first to go (yep), and more resolute damage such as caps popping, there's actually a MUCH more common failure point than the GPU die: The memory / VRAM. Part of my research and bigger 'talking points' if you will about the benefits of immersion cooling in the mining space was the additional resilience (near-immunity) to environmental degradation of hardware. GPUs, ASICs, and FPGAs primarily experience degradation from their environments in the form of corrosion at an extremely small scale. Of the myriad forms of corrosion, the most obvious was due to the humidity in the air, which found lovely nucleation points all over due to the fine dust so common in large-scale mining ops. And wouldn't you know it, the most sensitive component on the GPUs to this corrosion was the memory. Especially since in most designs the VRAM modules are so exposed to air (rarely being totally covered by a heat sink, and often being on both sides of the board), they would experience corrosion less visible than other spots on the PCB, but more unfortunately more damaging. You would find more *visually* obvious corrosion around power delivery components, especially the 12v connectors and caps, but these components were usually a fair bit more resilient than memory. Certain mining algorithms focus more on GPU (compute), and some more on memory (VRAM, both timings and effective clockspeed). In a lot of cases you could end up mining for months or years on a relatively "lax" algorithm and simply ignore or not encounter memory errors that would surely cripple or disable a GPU on another algorithm. In our (LTT viewers) case, we're interested in gaming. Unfortunately, we fall somewhere in the middle as far as compute and memory getting hit goes - a fair balance the GPU is obviously built around. Best case, a GPU encountered minimal humidity and/or minimal dust, and simply didn't corrode. Worst case, it got an unlucky dose of both and is failing spectacularly shortly after posting. More commonly however, you end up with only a slightly degraded GPU at a hefty discount that, after a good cleaning, is A-OK to game on. If the GPU is encountering memory errors at all, they're usually few enough that the GPU can actually work around them. (This is something we actually have to occasionally check and account for in larger, more finely-tuned mining ops; the error-rate that is.) -- A few comments discuss thermal cycling as a possible point of degradation for the silicon. A very real and potential problem for even modern hardware, but in our case (as a gamer looking to scoop a mining card), not really a risk in our case. As Linus found in the video, his tests were within margin of error between cards. A couple people in the comments posited that, under mining conditions, GPUs are largely kept at the same 'target' temperature. This is indeed accurate, and for the most part algorithms are largely constant-loads. It behooves the miners and those developing mining software to find a way to constantly load the GPUs. If you give them time to "chill out" for a moment, they become less stable (under mining overclocks/undervolts). Additionally, almost all mining programs and OSs set a target temperature that is maintained by fan speed updates. Also a lot of miners undervolt their cards so they're actually running lower power and a little easier. That said, 24/7/365 workloads of any form, regardless of mitigations, will still wear on a card harder than a gamer who plays, on average, a few or many times a week, for only a few to several hours a day. Especially in the conditions you experience in a mining farm. No air conditioning (so higher heat), dust, dirt, debris, humidity, etc. So we're far less concerned with thermal cycling and MUCH more worried about the awful environment they're operated in. And this brings us full circle to the degradation I went over above. -- Linus covers it well. Inspect the card if you can. Do your best to buy from places that have reasonable buyer protections in-place. If it seems too good to be true, it probably is. I've actually helped guide farms on their GPU resales and tried to help them clean, test, and - for AMD cards primarily - flash back to a more gamer-friendly vBIOS. On the other side of things, I've inspected large farms on behalf of potential purchasers and found farms in states of disrepair that ensured their cards would never run properly again - especially for a gamer.
We are having similar situation because of some big companies (already sued for false advertisement) saying - don't use your immune system, only buying our product will save you... But we can't really talk about that of course...
I think it really comes down to if the miner changed thermal pads or not. Most factory thermal pads suck. You might actually get a performance boost if the miner did this. If not it can be sketchy because mining tends to push vram temps into dangerous territory.
It’s it’s been mining for years without being turned off I wouldn’t worry because the temp probably stayed the same for years meaning the expansion or contraction wouldn’t be an issue and in fact could be in better shape than a card that sat unused for years in a room where temps vary drastically over time. I would buy a mining card all day if it’s clean and I know it was never shut down. It works right? Give me a deal!
My brother has been daily using a reference XFX HD 5770 since 2010.. I think he cleaned it once or twice at best. They should last a very, very long time.
Just bought a 3070ti from a miner for $410 USD about $100 off similar cards on the website with 6 months of usage and not any dust in photos should be receiving it later this week ill update!
Miners always undervolt/underclock so I’d say your getting a pretty damn good gpu instead of one that’s been gaming hitting max clock then ramping down consistently for a year or 2
@@hujimix you don’t come close to hitting max hashrates without undervolting and underclocking. Most miners are smart enough to know what they paid for.
I love how they just keep top of the line cards on the shelf in the box just to go obsolete rather than getting them into the hands of people that need them.
I’d hope they only keep some newer ones as backup and keep one of each of the older ones for tests like this but….yeah. Not to say they should just give ‘em away, nothing wrong with selling them as used. Better than hoarding if that’s the case.
I just won't buy from a Bitcoin miner out of principle. They caused the GPU shortage, leaving gamers stranded without a supply of graphics cards both new and old for several years. I'd feel disgusted to give my money to someone who caused that on top of the pandemic and lockdowns that nobody wanted to have to deal with in the first place. I put my morals before saving a couple of bucks
You could've just asked your mining friend to get all the average clocks and plot their %delta from their manufacturer rated clock speed, then plot that against the GPU's mining age, but instead you essentially gave him the ability to pick and choose which cards he was going to send you.
you know what, after seeing this exact comment copy pasted by about 700 different accounts, I have no choice but to go with tech host farms for all my online crypto-mining needs. Thank you, and god bless.
Eth 2.0 is coming very soon, and they're switching to Proof of Stake which doesn't need a GPU at all. So maybe that would mean more used GPUs would show up?
Proof of stake will be eliminating GPU mining on ethereum. The difficulty bomb for this hard fork will make the difficulty so high that Ethereum ASICs may even have a hard time. Right now some miners are switching over to other currently more profitable currencies as ASICs start to take over on the network.
One minor point, when you mention thermal expansion/contraction as a reason electronics with no moving parts wear out.... that doesn't apply to a GPU that is being run 24/7 because it doesn't experience the expansion/contraction cycle. It is like the old incandescent bulb that is turned on and left on so that it lasts longer than the ones that go on and off multiple times a day. Their are two main things that are likely to go bad over time from teh constant use... any capacitors that may exist on the card as capacitors are designed to last specific amount of usage, and the bearings in the cooling fans.
@@engenheirometaleiro2268 gpu's have target temperatures in mining programs that they maintain by adjusting the fan speed. Colder room = less fan. Hotter room= more fan. This allows the card to stay at a constant temperature ensuring it mines the most efficiently it can without stale shares.
@@engenheirometaleiro2268 A very very SMALL amount of it, compared to normal "shut the machine down daily" usage - even *if* the fan is set up for a constant speed.
Really? Your sponsor is Wargaming, while they are going through the worst PR shitstorm over promoting in-game gambling to children and facing comunity team walkouts?
This is such a weird video for me, there's the World of Tanks advertising, sure, but there's also the fact that, self-aware fire gfx aside, he's providing a platform for his friend's mining farm to advertise his used stock. It kinda seems fine overall, I guess, I'm sure used mining GPUs wouldn't be that bad a purchase, but something about it really isn't sitting well with me.
I literally got 4 RX580's for 110 each in September of 2019 (I sold prebuilt systems with used parts). Crypto proper crashed in 2019, people had until the 3000 series launch to cash in on it.
IK this is getting aged, but could you test VRAM performance? I don't think core clocks should be affected as core is not fully utilized for mining (like ETH) so miners would usually DOWNclock the CORE and OVERclock the VRAM as far as they can go. Careful and mindful ones probably take more care of the cards than gamers (temps, lower power draw = efficiency = profit), but since VRAM is the main affected component it's the one that may expose the signs of degradation or failure. More to that, many cards don't have sensors on memory so your best bet to get the VRAM temp is the hotspot, which is only estimately indicative. I was getting the same "GPU temps" with underclocked core (-502Mhz) and the memory range from -502Mhz up until +2000Mhz (where mining returns errors) for the test, which concerns as there should be a more heat due to higher clocks. Changing the core clock affects temps massively. Also, mining uses different algorithms working with VRAM unlike games, so it's possible to overclock much higher while in games you will get either artifcats or ECC at much lower offset. On top of running high clocks 24/7 would be interesting to see if there's really any dips with good care.
I thought mining doesn't affect the core at all instead it's heavy on vram. Also the fans might have degraded since miners usually run them at max or close to max speeds.
I usually go with whole gpu disassembly after first test run, cleaning thermal paste, removing thermal pads, take a zoomed look at the pcb for possible defects, check the gpu die if it really is what I was told it is, despite software saying it is that, then put all thermal stuff back, put gpu back together and update gpu's bios even if it's the same as current. Of course GTX 480, 660Ti and 970 probably weren't used for mining, I did see performance improvements (for 480 it was basically about 10m more before it was overheating) in few percent difference. 980Ti that I currently have was used for mining, but doing whole process gave me 15% performance boost (and +1% above marketed performance). Downside was that one of the fans started to get loose and wiggle enough to loose a fin and ramble even more, before eventually stopping working completely (I didn't do anything about it, except thinking I should do something, for 3 months so that's why it got so bad). Replaced with fan from other gpu in electronics store who repair stuff, because buying same fans where 1 is dead cost me 2$. Cut cables, soldered new working fan in it's place and works like a charm. And even looks like it wasn't changed, despite being from 1080Ti. Both MSi btw
Glad to know the RTX 2070 Gaming Z seems to perform well after mining, since i have the same and have started mining on it. PS: I get average 43MH/s on this card
Wow!! this video is just so wonderful . But cryptocurrency trading requires strategies of an expert to avoid unnecessary loss while trading. Mrs Jennifer has been my broker and I’m really proud of my financial status right now because of her. She’s really reliable and trustworthy.
It’s indeed wonderful seeing someone else talk about Mrs Jennifer here. I started trading with her after a friend recommended her to me, she’s really an expert in trading.
Mrs Jennifer has been a blessing to me. I was able to clear my masters degree fees through the profits i got from trading with her. I wish i had started with something more bigger.
I’m currently using Mrs Jennifer as my broker. She’s a pure definition of a reliable and genuine broker. Yes she charges 20% of the total profit. She assures you that you can’t loss your capital while trading with her. I like her for that.
interesting enough, the same card, from the same manufacter, from the same place and date, may have some differences when it comes to mining, since they most likely will have different memory types
I am not buying those because of what manufacturers say. I am not buying those because i am not going to reward miners for creating the GPU shortage and price hikes. I hope they get stuck with all the GPUs rotting under their asses.
@@MrTorch-p4z never say never. Crypto is not the only thing falling, almost an entire stock market has been down for a year, stocks hitting 52 week low left right and center, even the most formidable ones. But I'm pretty sure they'll come back up, the question is when, and that's probably when crypto trends will increase again. So yeah.
@@lazry3208 yeah but not gpus, pc gamers check frequencies constantly due to overclocking and extracting as much performance as possible, tldr : tht marketer is full of sht
