E-Pal has large and trustworthy communities! Use my link go.epal.gg/3ELnGDu or Invite code "techlens" for an additional $5 and find your gaming partner today! ⭐ Watch THIS Video Next ⭐ ruclips.net/video/NrqeMm0nY94/видео.html ⭐
Informative video, But 1. Shouldnt we start with FurMark first as it is an holistic test? and then go to the nuanced test as mentioned by you? 2. Also if a card is mined(without us being aware) but its under official warranty and we have the bill, Can we get a replacement(Within the 2-3yr warranty period it has remaining)?
Honestly, you can run them in whatever order you like 😊 that doesn't matter too much. Good question, I've heard stories of companies refusing the warranty for mining. But realistically, there's not going to be a whole lot they can do if you just say "it has never been mined on", even if you don't know. Best thing to get is a cash receipt from the buyer if possible. If they paid by cash, you have the original receipt and there's nothing linking someone else to the purchase. You're golden. Got an amazing deal on some memory a while back. 1 stick was completely defective, but they had the receipt. So I bought it and did the RMA myself and was super happy.
Considering the current pricing of electricity in my country, after running the full suite of tests you showcase, I would easily have spent the difference between used and new gpu prices in bills 🤣
I bought 2x 3090 Founders Editions that were mined on. I changed out the thermal pads and paste the day I got them. Dropped the VRAM from its normal 100C down to 80C. The Card was clean with no dust. With no adjustments, the clocks stay stable at 1950 Core. I have put at least 100 hours on one of them, and 500 hours on the other. Zero issues.
@@LanaaAmor I didn't touch any settings, replacing the Thermal Pads on the VRAM chips is what allowed the massive drop in temps. Both of my 3090s are still running solidly.
@@LanaaAmor There are plenty of guides available online. Recommend Gelid Extreme 1.5mm Thermal Pads, the Ultimate stuff was too hard. The idea is that stock pads have the VRAM running close to the limits of GDDR6X's thermal limits. The new pads bring that down comfortably within spec. This in turn would prolong the lifespan of the card. The 3090 has had a few cases of VRAM chips fail prematurely.
A while ago (when the mining was still peaking) i heard about a tool named MATS MODS. I remember that the tool checked the memory corruption in the graphics card. So in that way you can test which memory module is good or broken. (Which is a main component that is used while mining). And then, you could replace the memory (via microsoldering) and that’s it (i think).
MODS = Nvidia Modular diagnostic software is actually the main diagnosis program while MATS = Memory Automated Test System is the standalone version which only checks for memory errors and such
Great guide, apart from one thing. DO NOT just use FurMark willy-nilly as a normal lay-user. The application is meant for experienced overclockers who know to manage performance and temperatures. If you use it without the proper cooling, it is simply a frying simulator, where you need to turn it off before your GPU reaches 100 degrees C. I did a test where my GPU reached 110 C and then shut down. Thankfully no major damage, but I can't recommend this unnecessary step to anyone else. Can you run the newest games for hours on end whilst also having photoshop or blender open in the background? And all of this under 80 degrees C? Congratulations, you passed. Furmark is essentially a boasting benchmarking tool for measuring you know what replacements. The practical use for the average user is limited, when compared to other more reasonable ways of benchmarking. I can run Skyrim with 1500+ mods ENB 2K textures and everything else cranked to max with 115 variable FPS (I lock to 61 for stability though). Baldurs Gate 3 is a stable 144, and Elden ring max settings I get 95 FPS. I also use VRAM heavy tools in the background at times. For me there is no need for any further "stress testing" - these scenarios simply don't exist. Games that use 120 % of your clock speed should be illegal, and I haven't tried any either. No game engine starts out with a mission to destroy GPUs, in fact the opposite is the goal.
Woooow. this video is so full of usefull information delivered in the most clear and concise manner possible! :O neither dumbing down anything or assuming much prior knowledge. Pure gold, thank you! :D
This is an absolutely fantastic video. You are straight to the point, direct, no wasted words, very efficient with your communication and thoughts. Thankyou. I'm so grateful you're not a bloviated word-waster like so many others. Liked, Subscribed.
Thanks, man, for your wonderful guide. I purchased a second-hand (not sure if mined) 6700XT yesterday. The GPU has turned out fine and this has been a great help.
@@alexandercuyson1856 yeah it is. But 1 issue i have is that the screen is flickering in certain idle scenarios which is annoying. But in game and fullscreen videos/programs, it works just like normal. so 1 issue but its still working fine for my needs.
Thank you so much for teaching me how to graph the log data! I've been searching for a while how to do this, but your the first one to show how in one of these types of videos!
Personally, I don’t recommend running Furmark, except for short term testing. If you don’t know what you’re doing, don’t touch it at all. It’s more likely to explode your caps and cook the VRM, or pop a fuse on that card. It’s more of a power virus than a gaming workload. I’d recommend, and this isn’t free so I guess thats why he doesn’t recommend it, is 3D Mark stress tests. Just leave one of the Fire Strikes, or Time Spy running on a loop. Each test has different properties that push the VRAM, and pushes higher or lower frame rates. I’ve had very good results that are reliable with that.
Usually a slight memory overclock and test will reveal if the card was mined. The miners do that memory overclock in order to gain more same time underclocking and undervolting the GPU in order to get more efficiency for the power consumption. So the first clue is an instable memory or memory crash when you overclock it. The memory chips do get damaged and "older" with frequency increase and high temperature.
the best video about checking both used GPU i've seen. GJ - subscribed to you because of it. Both month ago a used 6700XT, will test it like you just did.
Always test any component prior to maintenance or modification. New cards these days, decent ones, not Dell or OEM reference garbage cards. Newer cards come with good paste and pads from the factory. You’re more likely to make things worse, you’re better off leaving it alone unless there’s an issue. Almost killed a brand new 3090 applying Kryonaut, only to discover there’s a good chance it came with Kryonaut from the factory. So I messed something up that was shorting the cooler off of the PCB, by over tightening, and got extraordinarily lucky it didn’t kill it. It just wouldn’t even recognize the card and the bios is like beeeeep x5 times, which often times indicate a bad connection or electrical issue. Had it happen again recently when a fan wasn’t plugged in completely. I did get a 1080 with the cooler not installed correctly from the factory. So that was a necessity
I purchased (by accident due to the ad changing daily) a 3080ti off Amazon. So far so good. Cards been preforming well and benchmarks above average but I'm constantly worried about it breaking.
A card that has been mind on for two years, is under a constant load at a steady rate so there is no thermal throttle they are generally undervalued and kept in dust free environments. So in most cases. In better condition than a card that has been played on heavy and overcooked for two years. It’s been proven many times that mining is less stress on the card by around 10% so it’s better to buy an ex mining card than an ex gaming card.
I purchased 3 rtx 3070's for 300 each. I'm happy with my savings and performance. My home is now equipped with 3 great gaming pc's for about 800 each, or the price of one new high end pc. If the card have made it through 18mo of full use. They'll be okay for gaming, apply new paste and ride.
Warranty sticker being removed just means they may have replaced the thermal paste. I personally replace the thermal paste with every new card i get. 😂
@@KeiGGR tf r *_u_* on? I said that a smart chip that would measure out the mileages would jack up the prices. Sure the HDDs may be cheap now but not with something expensive alone like this
I just bought my RTX TUFF 3070 2 months, from ebay, for 300 bucks, still in the original box, IT STILL HAD THE PROTECTIVE FILM on it! Pristine condition!
For a noob, this video is so darn deep. I've watched it almost 4 times now. Side note, my $25 RX ~~380~~ 580 (2304 shaders lol) is performing like a champ! And I'm breathing really easy now. Thanks,
@@lonewolf2150 I sold the gaming rig to a concrete company for CAD design, of all things. I gave them a one year warranty and haven't heard back, so... Bingo! Now I'm playing with a Zotac/Corn GTX 960 AliExpress special I bought off eBay. Because I can't seem to learn my lessons.
Mining requires underclock not overclock, specially the gpu. As for the memory depends on what coin you are mining, some need a bit overclock and some don't, and others need to lower memory so gpu can work better. I have been mining since long time and my cards are all the AMD from Sapphire and temps are below 50celsius. So why you trying to make some people get frightened? For example Raven coin you need to underclock the GPU by 20 to 25% less. It gives same hashrate as normal clock but consumes less power and gets cooler.
"ingame" benchmarks are not great for FPS orientation bcs they are more CPU and RAM dependable then most of synthetic 3dmarks.. where u can have dedicated tests just for gpu
Here in the US, those stickers are worthless. You buy it, its yours, and you have all the rights to take it apart for any reason. Manufacturers cannot deny service for breaking a sticker. The owner has to drive a screwdriver through the pcb or something for warranty to be voided. Also about fans, don't freak out if you buy a used card and the fans don't work. Clean/oil them because the owner could have been a smoker, and that gpu sitting around unused can allow the tar to congeal effectively acting like a glue.
Can someone not just print those small circular warranty void stickers? I mean how difficult can it really be....with sticker printing being so easy and cheap.
Fresh New Used GPU? How about take it from a line from the film, "Hello, My Name is Doris," and call it "Fresh Vintage?" Thanks for the tips, and also holding out for further price drops for the next couple of months.
To anyone who is considering buying a used mining gpu, some good ways to verify if a seller is legit before buying are checking if they are verified on their respective platform , seeing how much information they give about the card, and asking questions. I recently bought a 3080 on Mercari and it was completely legit.
@@prajit6483 How long they had it, what games they played, etc. You could even ask for benchmark scores if you really wanted to. There are a lot of ways to figure out if the card is in good condition. A lot of mining cards will be fine but some that are stored and cleaned improperly can be significantly worse from the get go or deteriorate rapidly.
you're insane. no way i'd spend more than an hour total testing stability, let alone a week like you're suggesting. if nothing crops up in an hour, it's unlikely that there's something wrong.
considering low understanding of nuances by average Joe ... often they end up blaming the card while the problem was in insufficient or poor quality PSU. and in some cases bad fans are replicable if pricing by seller was alright. stability or low fps are the biggest problems for GPU.
I definitely, well, I suspect I got a couple of mining cards. I assumed when I bought them they were mined on. It’s pretty obvious when a seller has multiple of one type of card. The Zotac 2080 was a blower style, a bit of dust in it, nothing crazy. It turned out even though the power limit was gimped from the factory, that card ran like a beast. My Dell 3080 on the other hand, it was near spotless. Obviously disassembled and modded for improved vram temps. And that card runs like crap! There’s nothing wrong with it, it’s just so pathetically underpowered, it can’t even maintain 1 volt under load. It works great, it’s just a very bad reference model. Wishing I had spent an extra bit of money and got that used FE. The point I wanted to mention was one of the fans bearings was on its way out. It would make this crazy rattling sound at around 60% fan speed. I liquid cooled it which helped immensely, but the VRAM still cooks for some reason. Crappy water block, Alphacool Eisblock, I took one look at it and was like, I haven’t seen a water block this poorly designed in almost 10 years. It was cheap, got the card and built the loop for under $600. I think it needs a better pump, I got a 360mm rad for $45, and a Chinesium pump and reservoir for $25. Works better than I thought, not as good as I’d hoped. Core temps around 50C, VRAM still in the 80’s though, which is HOT for liquid cooling.
the first thing to do is check how clean the card is, a dirtier card likely just a gameing used card, if it's almost spotless the first thing to do is just let it sit out for 3-4 days, to make sure the card wasn't washed with water. some of these miners will take a damn water hose to these cards to clean them.
So when running the OCCT 3D test, it is above the base clock for 99.98 percent of the time but has maybe four dips below the base clock throughout the test, is this normal or would that be sign of a faulty card?
I got a used 5600xt MSI gaming MX.. I did a fan test but there it was inconsistent in a different way.. the amount of increased RPM spikes is more than when it slows down ? is that a good sign or a bad sign or both or what is it ?
I need help I bought 6600xt on eBay once I installed amd software my PC won't boot but I have another 6600xt that one work fine what can I do to fix the other one?
@@ZeroRedz A stroke can cause damage to the cerebellum or basal ganglia in the brain, resulting in a tremor. Shaking and trembling can also be a warning sign of a future potential stroke.
Absolutely no need for testing, just don't buy a gpu from miner under any circumstances. Considering what they did, they deserve to lose as much money as possible. It's needed to teach them a valuable lesson.
@@outa7192 I already have 6800 xt, why I would need a new gpu? I bought it 2 weeks ago from reputable reseller with 2 year warranty. And for these scummy companies, they will lose quite a lot of money aswell, they already are losing. Miners were and still are the biggest enemies there could be, if you buy from them you basically tell them it's ok to do exactly the same thing next time around. They need to pay for what they did!
@@PanzerFaust1754 1. Not everyone is able to buy msrp. 2. That would cause a lot of e-waste. 3. The initial problem is the fact that these companies supplied cards to miners which contributed to the problem so if you really want to teach someone a lesson it would be these companies. They don't care who the consumer is as long as it line their pockets.
@@outa7192 1. I'm from Europe, if you would check our prices aren't even remotely close to msrp. 2. Not really. 3. Not quite, companies usually sold gpu to their distributors and what distributors did is the result we see today. So let get me this straight, instead of punishing miners who were and still are the problem you want to punish the companies?
Any clue why in "corectrl" under linux, my powercolor radeon RX5700xt reads 180w max? Is the card undervolted from the factory or is this just an erronious readout?
E-Pal has large and trustworthy communities! Use my link go.epal.gg/3ELnGDu or Invite code "techlens" for an additional $5 and find your gaming partner today!
⭐ Watch THIS Video Next ⭐ ruclips.net/video/NrqeMm0nY94/видео.html ⭐
Informative video, But
1. Shouldnt we start with FurMark first as it is an holistic test? and then go to the nuanced test as mentioned by you?
2. Also if a card is mined(without us being aware) but its under official warranty and we have the bill, Can we get a replacement(Within the 2-3yr warranty period it has remaining)?
Honestly, you can run them in whatever order you like 😊 that doesn't matter too much.
Good question, I've heard stories of companies refusing the warranty for mining. But realistically, there's not going to be a whole lot they can do if you just say "it has never been mined on", even if you don't know. Best thing to get is a cash receipt from the buyer if possible. If they paid by cash, you have the original receipt and there's nothing linking someone else to the purchase. You're golden.
Got an amazing deal on some memory a while back. 1 stick was completely defective, but they had the receipt. So I bought it and did the RMA myself and was super happy.
Bro I brought a prebuilt pc I need to check is it a mined one how can I check
Ppp
Considering the current pricing of electricity in my country, after running the full suite of tests you showcase, I would easily have spent the difference between used and new gpu prices in bills 🤣
hahahaha
Same
Absolutely incredible video. High quality content and could tell a ton of work went into this. Hope your channel continues growing!
I bought 2x 3090 Founders Editions that were mined on. I changed out the thermal pads and paste the day I got them. Dropped the VRAM from its normal 100C down to 80C. The Card was clean with no dust. With no adjustments, the clocks stay stable at 1950 Core. I have put at least 100 hours on one of them, and 500 hours on the other. Zero issues.
did the miner overclock it or did you underclock it? Also isn't screwing with the settings risky? pls give more tips
@@LanaaAmor I didn't touch any settings, replacing the Thermal Pads on the VRAM chips is what allowed the massive drop in temps. Both of my 3090s are still running solidly.
@@jakthebomb how do u replace it? does it help in performance or make the card last longer?
@@LanaaAmor There are plenty of guides available online. Recommend Gelid Extreme 1.5mm Thermal Pads, the Ultimate stuff was too hard. The idea is that stock pads have the VRAM running close to the limits of GDDR6X's thermal limits. The new pads bring that down comfortably within spec. This in turn would prolong the lifespan of the card. The 3090 has had a few cases of VRAM chips fail prematurely.
@@jakthebomb but isn't it the actual graphics chip that fails usually? Also is it beneficial on used mining cards like 1050ti ?
A day of furmark? Definition of overkill.
A day OR MORE, he said. lol
If the card wasn't fried yet, it will be lol.
A while ago (when the mining was still peaking) i heard about a tool named MATS MODS. I remember that the tool checked the memory corruption in the graphics card. So in that way you can test which memory module is good or broken. (Which is a main component that is used while mining). And then, you could replace the memory (via microsoldering) and that’s it (i think).
MODS = Nvidia Modular diagnostic software is actually the main diagnosis program while MATS = Memory Automated Test System is the standalone version which only checks for memory errors and such
@@MaxC_1 wow i didn't know that. Thanks for the info.
Great guide, apart from one thing. DO NOT just use FurMark willy-nilly as a normal lay-user. The application is meant for experienced overclockers who know to manage performance and temperatures. If you use it without the proper cooling, it is simply a frying simulator, where you need to turn it off before your GPU reaches 100 degrees C. I did a test where my GPU reached 110 C and then shut down. Thankfully no major damage, but I can't recommend this unnecessary step to anyone else. Can you run the newest games for hours on end whilst also having photoshop or blender open in the background? And all of this under 80 degrees C? Congratulations, you passed. Furmark is essentially a boasting benchmarking tool for measuring you know what replacements. The practical use for the average user is limited, when compared to other more reasonable ways of benchmarking. I can run Skyrim with 1500+ mods ENB 2K textures and everything else cranked to max with 115 variable FPS (I lock to 61 for stability though). Baldurs Gate 3 is a stable 144, and Elden ring max settings I get 95 FPS. I also use VRAM heavy tools in the background at times. For me there is no need for any further "stress testing" - these scenarios simply don't exist. Games that use 120 % of your clock speed should be illegal, and I haven't tried any either. No game engine starts out with a mission to destroy GPUs, in fact the opposite is the goal.
Woooow. this video is so full of usefull information delivered in the most clear and concise manner possible! :O neither dumbing down anything or assuming much prior knowledge. Pure gold, thank you! :D
This is an absolutely fantastic video. You are straight to the point, direct, no wasted words, very efficient with your communication and thoughts. Thankyou. I'm so grateful you're not a bloviated word-waster like so many others. Liked, Subscribed.
If you value your time and sanity, and just want to purchase a gpu that works as it should. Buy it brand new, or buy it from a friend you trust.
Thanks, man, for your wonderful guide. I purchased a second-hand (not sure if mined) 6700XT yesterday. The GPU has turned out fine and this has been a great help.
same. got it for around $310 and so far its been doing just fine. My card is for sure mined on tho since the guy was called Minebreakers.
Is it still working?
@@leifparon1561still working?
@@alexandercuyson1856 yeah it is. But 1 issue i have is that the screen is flickering in certain idle scenarios which is annoying. But in game and fullscreen videos/programs, it works just like normal.
so 1 issue but its still working fine for my needs.
How did you get through your GPUmemtest, it doesnt show openCL for my rx 6800
16:31 why? 💀💀
Excelent video btw! Thank you for all the detailed info
Another eBay purchase, and I'm watching this video again. Evergreen af. Thanks!
Thank you so much for teaching me how to graph the log data! I've been searching for a while how to do this, but your the first one to show how in one of these types of videos!
Personally, I don’t recommend running Furmark, except for short term testing. If you don’t know what you’re doing, don’t touch it at all. It’s more likely to explode your caps and cook the VRM, or pop a fuse on that card. It’s more of a power virus than a gaming workload. I’d recommend, and this isn’t free so I guess thats why he doesn’t recommend it, is 3D Mark stress tests. Just leave one of the Fire Strikes, or Time Spy running on a loop. Each test has different properties that push the VRAM, and pushes higher or lower frame rates. I’ve had very good results that are reliable with that.
Usually a slight memory overclock and test will reveal if the card was mined. The miners do that memory overclock in order to gain more same time underclocking and undervolting the GPU in order to get more efficiency for the power consumption. So the first clue is an instable memory or memory crash when you overclock it. The memory chips do get damaged and "older" with frequency increase and high temperature.
the best video about checking both used GPU i've seen. GJ - subscribed to you because of it. Both month ago a used 6700XT, will test it like you just did.
Well this was super comprehensive and just on time to test my (hopefully working) 3090 😁😁
Great video dude! my covered in dust 3080 arrives tomorrow, wish me luck lol
How did it turn out? Did it pass the tests? I’ve got a a used 3080ti coming in a few days
Any updates?
Always test any component prior to maintenance or modification. New cards these days, decent ones, not Dell or OEM reference garbage cards. Newer cards come with good paste and pads from the factory. You’re more likely to make things worse, you’re better off leaving it alone unless there’s an issue. Almost killed a brand new 3090 applying Kryonaut, only to discover there’s a good chance it came with Kryonaut from the factory. So I messed something up that was shorting the cooler off of the PCB, by over tightening, and got extraordinarily lucky it didn’t kill it. It just wouldn’t even recognize the card and the bios is like beeeeep x5 times, which often times indicate a bad connection or electrical issue. Had it happen again recently when a fan wasn’t plugged in completely. I did get a 1080 with the cooler not installed correctly from the factory. So that was a necessity
thank you for doing this piece,
thorough details on checking used GPUs :)
I purchased (by accident due to the ad changing daily) a 3080ti off Amazon. So far so good. Cards been preforming well and benchmarks above average but I'm constantly worried about it breaking.
A card that has been mind on for two years, is under a constant load at a steady rate so there is no thermal throttle they are generally undervalued and kept in dust free environments. So in most cases. In better condition than a card that has been played on heavy and overcooked for two years. It’s been proven many times that mining is less stress on the card by around 10% so it’s better to buy an ex mining card than an ex gaming card.
Exactly; Under constant load. The fans will go first. The memory can only handle so many reads and writes before those modules die.
I purchased 3 rtx 3070's for 300 each. I'm happy with my savings and performance. My home is now equipped with 3 great gaming pc's for about 800 each, or the price of one new high end pc. If the card have made it through 18mo of full use. They'll be okay for gaming, apply new paste and ride.
Are they still ok
Outstanding video.
Warranty sticker being removed just means they may have replaced the thermal paste. I personally replace the thermal paste with every new card i get. 😂
8:07 Booting Up section has a really annoying clicking sound (most likely a fan hitting on a cable)
12:25 fantasic and the fans in the background lol
To bad GPUs dont have a SMART. chip or something in the hardware like a Hard Drive has so you could see how many hours were on the card.
That'd jack the prices even further since those chips alone are likely worth another gpu
@@ItsDaKoolaidDude Tf u on? Hard drives have had them for ages and are not even worth 50 bucks now
@@KeiGGR tf r *_u_* on? I said that a smart chip that would measure out the mileages would jack up the prices. Sure the HDDs may be cheap now but not with something expensive alone like this
I just bought my RTX TUFF 3070 2 months, from ebay, for 300 bucks, still in the original box, IT STILL HAD THE PROTECTIVE FILM on it! Pristine condition!
When you "get" your card it's too late to test.
depends where you buy it from
For a noob, this video is so darn deep. I've watched it almost 4 times now.
Side note, my $25 RX ~~380~~ 580 (2304 shaders lol) is performing like a champ! And I'm breathing really easy now. Thanks,
380? I've never heard of this card
@@AhmedFahim16 RX 580, you're right. Thanks.
@@LesCalvin3 any updates? Still working?
@@lonewolf2150 I sold the gaming rig to a concrete company for CAD design, of all things. I gave them a one year warranty and haven't heard back, so... Bingo!
Now I'm playing with a Zotac/Corn GTX 960 AliExpress special I bought off eBay. Because I can't seem to learn my lessons.
@@lonewolf2150Sold the rig for 3D work. Last I heard, they were very happy with it.
Mining requires underclock not overclock, specially the gpu. As for the memory depends on what coin you are mining, some need a bit overclock and some don't, and others need to lower memory so gpu can work better. I have been mining since long time and my cards are all the AMD from Sapphire and temps are below 50celsius. So why you trying to make some people get frightened? For example Raven coin you need to underclock the GPU by 20 to 25% less. It gives same hashrate as normal clock but consumes less power and gets cooler.
"ingame" benchmarks are not great for FPS orientation bcs they are more CPU and RAM dependable then most of synthetic 3dmarks.. where u can have dedicated tests just for gpu
I can't even adjust my game settings, so good luck on me doing all these tests.
Here in the US, those stickers are worthless. You buy it, its yours, and you have all the rights to take it apart for any reason. Manufacturers cannot deny service for breaking a sticker. The owner has to drive a screwdriver through the pcb or something for warranty to be voided.
Also about fans, don't freak out if you buy a used card and the fans don't work. Clean/oil them because the owner could have been a smoker, and that gpu sitting around unused can allow the tar to congeal effectively acting like a glue.
Can someone not just print those small circular warranty void stickers?
I mean how difficult can it really be....with sticker printing being so easy and cheap.
Fresh New Used GPU? How about take it from a line from the film, "Hello, My Name is Doris," and call it "Fresh Vintage?" Thanks for the tips, and also holding out for further price drops for the next couple of months.
Fresh Vintage, love it!
To anyone who is considering buying a used mining gpu, some good ways to verify if a seller is legit before buying are checking if they are verified on their respective platform , seeing how much information they give about the card, and asking questions. I recently bought a 3080 on Mercari and it was completely legit.
what are questions you asked them
@@prajit6483 How long they had it, what games they played, etc. You could even ask for benchmark scores if you really wanted to. There are a lot of ways to figure out if the card is in good condition. A lot of mining cards will be fine but some that are stored and cleaned improperly can be significantly worse from the get go or deteriorate rapidly.
@@prajit6483prob a benchmark, a picture with his name on a sticky note with the graphics card or something like that
I get my RTX 3090 24gb used for only $500 and its awesome ex mining card, no corrotion and still in like new condition
Is it still alive
Did u buy yourself? Saw u commenting like hell 😂@@vainl4703
you're insane. no way i'd spend more than an hour total testing stability, let alone a week like you're suggesting. if nothing crops up in an hour, it's unlikely that there's something wrong.
considering low understanding of nuances by average Joe ... often they end up blaming the card while the problem was in insufficient or poor quality PSU. and in some cases bad fans are replicable if pricing by seller was alright. stability or low fps are the biggest problems for GPU.
Found a 6800 XT through Amazon Refurbished going for around $380. Should I risk it?
I definitely, well, I suspect I got a couple of mining cards. I assumed when I bought them they were mined on. It’s pretty obvious when a seller has multiple of one type of card. The Zotac 2080 was a blower style, a bit of dust in it, nothing crazy. It turned out even though the power limit was gimped from the factory, that card ran like a beast. My Dell 3080 on the other hand, it was near spotless. Obviously disassembled and modded for improved vram temps. And that card runs like crap! There’s nothing wrong with it, it’s just so pathetically underpowered, it can’t even maintain 1 volt under load. It works great, it’s just a very bad reference model. Wishing I had spent an extra bit of money and got that used FE. The point I wanted to mention was one of the fans bearings was on its way out. It would make this crazy rattling sound at around 60% fan speed. I liquid cooled it which helped immensely, but the VRAM still cooks for some reason. Crappy water block, Alphacool Eisblock, I took one look at it and was like, I haven’t seen a water block this poorly designed in almost 10 years. It was cheap, got the card and built the loop for under $600. I think it needs a better pump, I got a 360mm rad for $45, and a Chinesium pump and reservoir for $25. Works better than I thought, not as good as I’d hoped. Core temps around 50C, VRAM still in the 80’s though, which is HOT for liquid cooling.
GPUMemTest doesn't work on AMD GPUs.
the first thing to do is check how clean the card is, a dirtier card likely just a gameing used card, if it's almost spotless the first thing to do is just let it sit out for 3-4 days, to make sure the card wasn't washed with water. some of these miners will take a damn water hose to these cards to clean them.
Problem here is, I need to test 7 cards... this all takes way too much time :(
So when running the OCCT 3D test, it is above the base clock for 99.98 percent of the time but has maybe four dips below the base clock throughout the test, is this normal or would that be sign of a faulty card?
Extremely helpfull video. Thank you! You deserve more subscribers then 32k. GL!
I got a used 5600xt MSI gaming MX.. I did a fan test but there it was inconsistent in a different way.. the amount of increased RPM spikes is more than when it slows down ? is that a good sign or a bad sign or both or what is it ?
I couldn't find memory Junction? Why is it not on OCCT?
Hmmmmmmmm, should've put some of these in my video
Im getting an error on gpumemtest saying no OpenCL found on my rx 6800 xt, i tried ddu and installed new drivers and i have enabled inommo in Bois
All good, but there's no memory junction temperature anywhere.
I was looking for it too.
There is only one registered gpu fan in my hwinfo64 but all of my fans were working, why is that happening?
@bruh ohh i see, so theres no way to check if other fans were bad or good hehe. Ty
Fan tas tic fan tes ting
I just bought the RX 6800 and tried testing the Gpumemtest, it shows no openCL any fixes?
Nice video ❤
I'm getting a GTX 1660 Super!
My 3070 died after 9 months i did all this before i bought it. Only buy used if they got warranty in it.. otherwise it's just a gamble of course.
My 3080 works perfectlyl, I bought it after mining in 2022.
I got lucky I guess.
GTX 285 is 2008 not 2013 hehe
What about if memory doesnt show in GPUZ and theres no bios?
Isn't GpuMemTest only for Nvidia?
gpumemtest doesnt see my RX6800 any advice ?
Use the OCCT VRAM test instead.
I ended up using OCCT and everything else you suggested and tye mining card I purchased a rx6800 works like new thanks
8:45 Whats that beautiful case?
(Late to the party) looks like a lancool 3 mesh from lian li
How do you get these graphs?
This is the most overkill video ever lmao i wouldn't even do this much testing on a 50k used bmw let alone a 300 bucks gpu😂
😂😂😂😂
I need help I bought 6600xt on eBay once I installed amd software my PC won't boot but I have another 6600xt that one work fine what can I do to fix the other one?
Yo did you figure it out?
Sell it as a fake working gpu on ebay on a brand new account
Clear CMOS, and set pcie to gen 3 manually in bios
Great video , you are Manuel Neuer from wish 😂
Dont waste your time here. Not usefull video at all. Just do Mats, 3dmark, valley and if everything is pass you will be fine.
Why are your hands are shaking?
probably he''s nervous..
@@ZeroRedz idk man, Im more worried it caused by light stroke
@@anigame8916 wat?
@@ZeroRedz A stroke can cause damage to the cerebellum or basal ganglia in the brain, resulting in a tremor. Shaking and trembling can also be a warning sign of a future potential stroke.
Nice video but you torture so much a card it actually dies before you could use it /jk
27 miner be like:
Absolutely no need for testing, just don't buy a gpu from miner under any circumstances. Considering what they did, they deserve to lose as much money as possible. It's needed to teach them a valuable lesson.
Well then I guess you won’t buy gpus from Nvidia or Amd as they sold them directly to miners. Go teach these tech giants a lesson.
@@outa7192 I already have 6800 xt, why I would need a new gpu? I bought it 2 weeks ago from reputable reseller with 2 year warranty.
And for these scummy companies, they will lose quite a lot of money aswell, they already are losing. Miners were and still are the biggest enemies there could be, if you buy from them you basically tell them it's ok to do exactly the same thing next time around. They need to pay for what they did!
@Giooo don't buy used? The easiest way to avoid mined card.
@@PanzerFaust1754 1. Not everyone is able to buy msrp. 2. That would cause a lot of e-waste. 3. The initial problem is the fact that these companies supplied cards to miners which contributed to the problem so if you really want to teach someone a lesson it would be these companies. They don't care who the consumer is as long as it line their pockets.
@@outa7192 1. I'm from Europe, if you would check our prices aren't even remotely close to msrp.
2. Not really.
3. Not quite, companies usually sold gpu to their distributors and what distributors did is the result we see today. So let get me this straight, instead of punishing miners who were and still are the problem you want to punish the companies?
I never buy used GPU's.
Any clue why in "corectrl" under linux, my powercolor radeon RX5700xt reads 180w max?
Is the card undervolted from the factory or is this just an erronious readout?
My GPU Fans have different max speeds, with a difference of about 105-115 RPM.
Is that problematic?
Should be no isseu unless rpm also isn't stable.
I bought a 1060 7gb temps are usually 45 to 50 idle does that sound ok with a week to send it back
You should really only be concerned with temps at full load
Luvly boy x