Here's Why My RTX Graphics Card FAILED... (with proof - 2080 Ti)
Вставка
- Опубліковано 26 січ 2019
- With reports all over the internet of higher than normal failure rates for RTX 2000 series Graphics Cards (including RTX 2080 Ti, 2080 and 2070), I heard rumors of it being specifically related to the Micron GDDR6 Memory, though in this instance my RTX 2080 Ti from Aorus started randomly artefacting and crashing. I then investigated the issue and found it to indeed be the memory.
Current prices of RTX Cards: amzn.to/2MxpYJd
💯Merch - www.redbubble.com/shop/techyescity 💯
❤️Support Tech YES City Directly: / techyescity ❤️
💻Check out my Video Editing PC Here - amzn.to/2oFKlrb 💻
💰Cheap CDKeys - TYC3 for Games or TYCSK for Win 10 Pro for up to 12% Off - www.vip-scdkey.com/software/m...
💰
-------------------------------------------------------------------------------------------------
Music Provided by either: monstercat, chukibeats, epidemicsound, vlognocopyright, audio library, no copyright sounds, or bassrebels
#GDDR6 #ComputerProblems #IGotToHelpPrime - Навчання та стиль
manufacturers need to spend less time spent on RGB and more on quality control
This is more on Micron than on the GPU brands. They're expected to have stress tested their own chips before supplying them to the board manufacturers.
But rgb +5 fps!!!
This is not the partners fault but the memory makers issue...
the partners just replace the cooler on the reference cards and add various other things like RGB, better fans, or a bigger cooler. the FE cards don't have RGB yet they've failed too. its more down to the bad memory from micron, not cause more time was spent on RGB. you sound like a jackass.
@@ERMMM420 one question, dont they also provide brand/card specific firmware? Not just to help identify it in software but also to run all the bells and whistles (had a 980ti k|ngp|n...redicoulous overpriced but nice to have nonetheless...including 3 unique firmware(s) with a switch to choose which one was being used)
Just in time as my 2080 Ti died :D
Emanuele Does NVIDIA give refund?
@@allansh828 I think only in the first 2 weeks, still need to have them get back to me about my RMA.
2080 ti has not been launched yet..
So now you only have 1 kidnye 🤔🧐
@@babitasaha6655 what?
If i had this card and it died on me i would go down with it.
Well hopefully you'd be covered under warranty. But yeah, I'd definitely feel like dying if I couldn't get a $1200 graphics card replaced.
My Zotac did die after one month and I had the luxury of paying for RMA shipping
@@MrVuckFiacom hopefully he has life insurance boiii
@@MrVuckFiacom my msi rtx 2070 armor died after 2 month and nobody refund me anything
@@abrahamlaerion fuck
"you can turn a negative into a positive, in that you can keep that memory slider down there, and then one month before your warren....beeeeeep...." so funny but so true.
That gives me nightmares.
Best editing I've seen on the internet for a long time :D
The card failed because you called it a gravis card.
and also, grabbes card.
It was putting out awesome sound though, that gravis card 😂
Amulya kumar singh ohhh, graphics, lol
hahaha
I thought he said travis cards
This is terrible! Very bad quality control, this is yet another indication the RTX series launch was rushed.
It's not surprising nvidia needs to keep competition fierce. This is due to amd becoming much better an closing the gap of the gpu market. Especially with the 7nm card coming. If amd has a card that does 2080 speeds or more for a 100 dollars yes. Nvidia will have to price drop.
@@nickbrown8729 Not only that, but for the price increass over the last 2 generations you can expect quality improvemeng, not degredation
@@nickbrown8729 AMD/ATI Has been competitive before too. There is no way AMD will make a 100 dollar card that equals a 2080 in performance.
@@jdjeasy Quality has not improved. I am sure nVidias margins did tho.
@Brad Viviviyal
That’s completely absurd to say. 3 yrs old Pascal card? I suppose RTX are closing the gap with a 3 yrs old Pascal card as well?
(With all honesty, that jokes you made lack any substance to it, it’s not even appeal to AMD fanboys).
Just a quick note , if you want to see what memory you have on your card you dont need to take it apart , use GPUZ it will tell you exactly what memory you have on your 20 series card , i used it to confirm what was on my Gigabyte 2080 before overclocking it & luckily for me it was Samsung 😀
not all micron's have a problem
Only micron that begins with an 8XXXX
That's why you don't buy day one
is it OK now 2 mths later?
@@stevethea5250 it was 3 months when you send that message
@@Mallagen it was 2 months 3 weeks ago
@@stevethea5250 and 4 weeks is 1month.
@@Mallagen 3 weeks ago was his comment 2 months
My argument would be.If i paid $1200 for a card it must work.I don't want to declock it just to get it running that's not what people pay for.I am already disappointed in the rtx on performance now declock as well :D
Hell, I'd say you have low standards, even. I f I paid 1200$ I'd expect 2 new working top-tier GPUs. That's how it used to work.
@Immense Data
Time’s changed... I remember my new NVidia cards back then worked out of the box flawlessly.
Pierré Grobbelaar not a valid point. Anything regardless of price may have problems. Of course we all want our equipment to work regardless.
Harald Berg
I still find it valid. He went out of his way & paid a premium price for that GPU, it gotta work out of the box right away for a high-end card. If you have to go through all of the hassle just for said GPU to work properly, it’s not ok, especially for NVidia GPU’s standard.
Samuels Wander I don’t disagree with the point that we except it to work. My point is that price have nothing to do with it. We all expect a GPU to work regardless if you buy it for $450 or $1200. Furthermore I am certain that both AMD and Nvidia have high quality standards. the same issue can be said of cars. BMW, Mercedes and other manufacturers have high quality standards. Yet from time to time they all have issues and they are repaired or recalled for repairs.
I work at an RMA centre and boy you wont believe the number of RTX cards thats been sent in for artifacting and crashing. It got so bad to the point we have to stop selling the RTX cards until the problem is fixed
Wow!
Good
@@markoglazersout611 That is good
Is the rtx 2060 one of them or only 2070, 2080 and 2080ti?
@@Kennyice23 i havent received any RMA for the 2060 yet
11:15 that’s what we do in Sweden. My cousin he gets a rtx 2080ti and he need to under clock his gpu to make sure his gpu not crash in 100% load.
Would be better to just ask for a replacement. Consumer laws in EU cover this kind of crap. prove it crashes in full load at stock with a video, and if they refuse to reimburse or change for a functioning one, sue them. Very easy win.
@@Enderplays12 i'm thinking of this
Ty soooooo much for this, I had exactly the same issues with my RTX Card
IT JUST WORKS!
@Dude Sensitive Ass REEEEE
It just breaks.
lol
yup it just works!...... for a few hours :D
Its amazing how many times he said that on stage lol
I swear he says "gravis" card ! :P
Well, Google's autosubtitles agree
So yeah, a new unknown 'artifact' on these ray tracers is the pronunciation. They are officially Gravis Adapters.
Gimme that Gravis UltraSound.
I think it is "grathiths"
Definitely yes
Excellent advice, I'll be checking your channel more often ^^
My Zotac RTX 2080 TI AMP (bought in November) had a catastrophic failure 2 weeks ago, the store replaced it with a Aorus RTX 2080 TI today.
Just happened to me this past week. Bought Zotac RTX 2080 AMP and failed
Do you think the newer batches are fixed? I really want to get one but not if it does this crap
Why not use GPU-Z to identify if it's Micron or Samsung memory ? it's faster :) The same problems have also been reported with Samsung memory (HardOCP).
He is bored 😂
Im on the second RMA now. Second card came with Samsung and yet failed in just 3 days. Now it shows error 43 in device manager. And no amount of "clean drivers" or even "clean OS install" can fix it. Luckily i have my poorly repaired 1080Ti with wires shunting the burned places after a liquid spill on my watercooling rig. But even in this frankenstein state it works like a charm. No problems for 1.5 year whatsoever. It just works :)
He wanted to see the board
could just be a bad design across all GDDR6. there may be a revision of GDDR6 in the pipes
@@PanaehaliTut my cooling system did the same, but leaked on a single RAM chip for my 980Ti, managed to clean up the residue and RMA that card, but when it came down to purchasing a new GPU, opted to buy a 1080Ti as i was not convinced the new RTX cards were going to be "stable" as what happens with most "NEW AND IMPROVED" computer components.
Im sooo glad I stuck with my Gtx 1080! :O
MetalGamer666 1080 is legendary card
Glad I didn't upgrade my 1050🤔
@@pewnice92 1050 is kinda slow though :D
Same my GTX 1080 just keeps going.
My 750ti don t want to die
my evga 980ti i got 3 years ago got a memory fail (micron) back then you could still look it up with software. and i got a gtx 1070 as a replacement,
Same here with a EVGA XC Ultra. Started artifacting on Witcher 3 and Pubg.
-200 on memory. Another 4 days. Same artifacts.
-500 on memory.. Another 3 days good... aaaand artifacts.
-99999 (slider all the way to the left).. Crashing all over the place.
Waiting for my replacement card now.
Hey did you get a replacement? And how it is now? Is everything working?
I got your point about warranty lol
I was like:_ WTF thats illegal :D but... he cut that part out... WE ALL KNOW what he say :D
@@VirusXAX I too understood that
@@VirusXAX why is that illegal? They have to replace the card. If sooner or later, why should it make a difference?
@@Gromic2k at least in my country you must report any issue within 30 days of noticing it
@@jarnom85 but how do they know you haven't only just discovered such problems?
Insane. The most expensive "gaming" GPU and it has quality issues! Unreal !
Well, only a few of them have quality issues... classic case of defective product... warranty takes care of it. These defective cards do not represent the official, functional product. :P
@@MyouKyuubi not a classic case of defective problem Nvidia should have spotted this issue before it launcheded
@@MisterCOM No, they couldn't, because the RTX cards were selling so fast they sold out instantly as their product released, INSTANTLY, in addition, now they had a bunch of orders to fill out, and no RTX cards ready to send them, so, they had to speed up production, which means a higher chance of things going wrong.
I had to wait months to get my rtx card, and it came fresh from the fucking factory. :P
@@MyouKyuubi how does any of that explain why they couldn't have tested them before launch?
@@MisterCOM Exactly!
I'm on my second third replacement card for a build I bought from Overclockers and it still powercycles in games on medium load (Mordhau, pubg, etc). Its like having a Ferrari that bursts into flames if you go over 20 miles an hour.
Interesting video!
One thing I want to add is the importance of using separate PCIe power cables for each power connector rather than using a single split cable if you have a modern PSU working in Multi Rail OCP mode.
Failing to do this could result in sudden hard resets of your system triggered by the PSU to protect your system. A feature rather than a problem but rarely feels like it when your PC resets out of the blue...
You can also avoid this by switching your PSU to Single Ray OCP mode. That however might result in your PC stating to smell melted cables and/or PCIe power connectors or worse in the long run.
Appalling crap from Nvidia. You pay as much for a graphics card as most people would spend on a whole computer, and you get to do the beta testing for them. Utter joke.
@Adam Saint it’s traditional to nVidia at this point
@@pacifistduck9615 *For any company*, fixed that for you, any company that puts out a new product that is faulty will get backlash..
its a tad fucked considering the rrp on these cards
This is kind of ridiculous. It's almost as if they never tested the longevitiy of the memory/other parts in the graphics card xD
they don't spend money and time on quality control because they knew people are going to buy any shit if the marketing is right.
*"Bleeding edge hardware"* with *bleeding memory chips* lol
*Bleeding edge wallet*
*Profusely bleeding bank account*, @@grimgoreironhide9985.
Main reason why I wait a few months, companies have taken so many shortcuts in recent years that I would not be shocked to see them use a nail gun to install parts.
Quit giving these companies valid cost cutting ideas! Lol
Good i still have my gtx 1070ti that i also overclocked. Was thinking about changing to 2080 half a year ago but i did not change because of this problem and it still exists wow ...
Gigabyte 2080 dead after 4 months. Scan uk refuse to refund though
Don't Scan give 12 months warranty as standard on components ? They should at least offer to replace it.
TubA222 RMA.
@@TubA222 card actually came with 4 years gigabyte warranty, however to claim that, it may take a longer time. I have decided to send it back to the shop. They sent bad card to the manufacturer who has confirmed the fault. It has been 3 week since them and finally after 5 emails they have sent something back to me. They still not responding to my emails with questions is it a replacement or refurbished product. Every time I mantioned the refund I had no reply back. Very rude and unprofessional. Will use amazon next time
@@logicmadness1752 If you do lots of shopping on amazon then worth it because you can score on lots of little discounts which all add up ... but I find prices are always better on newegg (for the stuff I buy, anyhow).
@logic madness - Damn that's unfortunate. You should phone them and ask to either escalate the issue or ask to speak to Liz Nicholas, as I believe she handles all customer service failures.
I liked the end there. “You can wait till warran...”, now back to our regularly scheduled program. Good bit of info Bryan. (:
Room setup is looking nicer and nicer mate. Glad to see you on the up and up!
I just bought and install an RTX 2070 last Saturday. I hope it doesn’t crap out.
underclock all your cards a bit to prevent shit like this
It still looks very much like a power sequencing issue to me. There are multiple power supplies on the card that must start up in the specific order at power on time.
It's possible that incorrect power sequencing results in a temporary 'out of spec' voltage condition on the RAM, causing incremental damage to the silicon every time the machine (actually the card) is powered on. With the various power management modes of the card, it's possible that same applies to those modes, and not just on/off power cycling.
We had a fair number of voltages and boost/power states on graphics cards for years now. What specifically would make you think that the sequences are at fault here ?
@@Kasapin5033 We've had a fair number of voltages and boost/power states on a lot more than just graphics cards for years now, but that's not the point. Sometimes a bad design makes it to market. I've seen the issue pop up multiple times on various designs and the symptoms of gradual degradation match what's going on with the RTX cards.
There are multiple comments suggesting that both Samsung and Micron equipped cards have displayed the fault. We can rule out bad RAM.
There have been thermal tests done under load showing that temperature was not the cause.
So, something non thermal on the cards is killing good RAM chips, that leaves a voltage issue being the most likely culprit.
This might be the problem. It could be just a power feed failing because the card is stable under lower clock speed . Normally lower voltage will cause instability in clocks speeds.
@@MickyMouseLimited The symptoms would certainly match the case presented in the video, but I think both the general instability/crashing and the space invader artifact issues are from the same cause (it just depends on whatever random damage happens to the RAM - causes different effects). A failing power feed cant account for the space invaders.
We really need to see if the card degrades further.
A good test would be to find the actual failure frequency, and stress test it just below that, see how long the continuing degradation takes to start affecting the card at the new frequency. Then re-adjust the frequency down to find the new stable point, rinse and repeat.
From this information, we could estimate when (if) the degradation will start causing trouble for RTX2060 owners as well.
It could be a problem with the new GDDR6 memory controller
*Tech Yes Warranty Tips*
I have the exact same problems here - artifacting, bsods, crashes causing failure to boot ultimately. Degradation progressed over a 6 month period until it became barely useable. Am now RMA'ing. However, mine has the Samsung GDDR6 memory rather than the Micron. Card is an nVidia 2080 Ti manufactured by Palit - GamingPro.
-The way it's meant to be played!-
Now that you said that, Triplex5014, people will quickly turn "it's" into "you're" for "The way you're meant to be played!". Seems rather appropriate nowadays...
Just buy it!
*I bought the rtx 2080ti so i can have realistic lighting in minecraft.*
I just care for people with blured vision
can u put avideo
@@yarox3632or people with failing RTX cards ;-)
But what about Spider Solitaire!
You dont even own a 1050
You can use GPUZ to pull the memory mfg off the board without tearing it down.
only if the ID's and calls are implemented, which may not be the case yet.
My MSI 2080 TI Gaming X Trio has Micron, at least in theory, according to GPUZ. I've no plans on taking it apart to verify but my expectation is it's correct. Which is why I suggested this as an alternative method of discovery. FWIW the card was unstable initially, but after flashing with a 'unofficial' BIOS that raised the wattage limits from 330 to 405, and using MSI afterburner with a custom, more aggressive fan speed profile to keep things cooler, it's been stable. However, if ambient gets over 80 or so it will occasionally still have issues. For anyone considering overclocking it further, I suggest seriously considering water cooling.
Yep I used this and turns out my RTX 2080 Ti uses Micron GDDR6 great
@@ZolaKluke same, running great right now though. Good OC tho I haven't bumped the memory too much in the beginning and after seeing mine also rocks micron memory, I decided to keep the memory as is - just be on the safe side
@@Seventschiko Yeah I've never OC'd my RTX card, though I could do it easily since I've got a gigabyte windforce 3x model
I wonder if the people making the micron memory chips are overclocking them by default to make them seem more impressive to buyers but they're actually only stable at 13ghz.
Whoa this is exactly what mine Zotac 2080 Ti Blower was doing... I was worried it was my mobo or PSU. Thanks for the video. It went back last week.
Just an idea but i don't know for sure but could it be the Memory controller in the Silicon (Turing Architecture) that is Failing in some cards? As some people are having the same issue with Samsung Memory as well.
Trust me folks gravis cards are the future :)
11:11 holy crap, you got me there HAHAHA
Yeah! Best part of this video. Too bad he had to make that edit... for monetization purposes. Lol
I think ill pass this generation. My Evga 1070 FTW 1.0 with the Acx 3.0 cooling solution had a fatal design flaw that caused inadequate heat transfer away from the VRMs, causing fires at worst and a rapid death at best. People had melted their memory modules and even the PCB itself. Got it replaced with a card that included thermal pads in the hotspots, and I now feel lucky to have one of the best value cards.
Wassup Brian, yo videos are the best!
I thought after watching years of Aussie shows I learned the accent.
I thought he was saying Gravis card instead of graphics card and looked up the Gravis brand name lol.
Well I just ate at an Outback Steakhouse this week but apparently I'm no Aussie!
i worked in a furniture company. 80% of our products were for export. we had 20-21 different products. when i made a product and it had a fault in the wood, it was repaired. if it was a paint fault, it was re-painted. i had a drawer with stickers of each company that was buying from us. Furniture that had to be repaired, was recieving a sticker for a company in my country or any country near us, in south east europe. cheaper products. if it had a paint problem and had to be re-done , was recieving an export sticker together with the other ones. The best ones that came out perfect were the only that got IKEA and Leroy Merlin stickers. i worked at 3 other factories for different things, gear boxes, glass fiber kids playground toys ,etc. all had same thing. products were labeled for different companies/customers depending on quality. i hope that tells you something.. :)
This is so interesting your first person I've seen fix these out of 4 channels.
I had my msi ventus 2080ti die after a week, exactly how it does in the video, artifacting while in any 3d task. My second one is an evga black which died after a month, this one however crashes after a minute a so even while in Windows. Rma for it got approved while watching this vid. The card is great while it works, but definitely would not recommend it until the issues are resolved.
How did you go about getting it RMA’d? I have a ASUS dual oc 2080ti and it has been having issues I don’t know whether to go through Asus or Nvidia and do you think I could get a refund ?
And if you took the refund and bought a water cooled Vega 64 you'd have a card that works with no issues and performs close enough.
I want top tier performance with cutting edge tech so that’s why I buy Nvidia
I had Micron memory in mine. Played BF5 for 3 hours with RTX On, then nothing but BSOD. Tried the memory slider trick, still died. NVIDIA and local store replaced it no worries.
GPU-Z did tell me it was Micron!
I've went through two bads 2080ti's last month. One DOA, artifacting and BSOD'ing Strix and a EVGA with a short/fault in one of the fans.
My first XC Ultra fried when I was playing FH4 and it tried to make me play Spaced Invaders. It had very little artifacting before hand and was unrecoverable. My replacement XC Ultra was artifacting heavily so I replaced it with an FTW3 Ultra and it hasn't had a problem at all. Both the previous cards had Micron memory and the FTW3 has Samsung.
Have had my Evga XC Ultra 2080 since Sept 2018, I replaced the TIM with Kyronaut a few weeks after I purchased it. It has Micron memory. *Crosses Fingers*
go try some RTX on. Micron shows issues fast with it on ultra in bfV
Unfortunate about the Micron. My EVGA XC Gaming has Samsung memory. Strange how the more expensive Ultra has worse memory
@@hxhgonkill My XC Ultra has Samsung. It was random early on now most should be Samsung
I would have returned that just to be safe. Next thing you know it's going to fail 1 day outside warranty and you are going to be left with a $700 brick. No thank you.
My EVGA Ultra XC 2080 has Micron been almost 2 months have never had problem and artifacting in games haven't overclocked it.
I don't need to RMA without having issues so yea.
I just upgraded from my founder edition 1070 that has Samsung but my FE 1070 were hot at 80C in during gaming had to sell.
And this is why kids, you don't buy anthing that releases in the first month.
including games
Evidenced more by Fallout 76
Reminds me of the time I had a water cooled Pentium 4 and the water block on the southbridge cracked and started leaking into the AGP slot. Really strange artifacts followed by crashes.
Thanks to your tip I was able to use my Aorus 2070 Extreme card again. I got artifacts, lockups, blank screens and random reboots before lowering the mem clock. Now is running slow but stable. It took me 8 hours to boot my PC yesterday... My card has Samsung memory on it according to GPUZ.
This just happened to me and I’m so bummed out.
Never buy a new artifision grafik card when it first comes out, always wait for the next series in line as bugs and issues will most likely be fixed and also the second series of that type of cards in this case ''RTX'' will come out alot sooner than we are used to with other cards and will also be alot cheaper and a bigger bang for you bucks 👍
Or it could be twice as expensive. Dont keep your hopes up until they get some competition
It's called the "bleeding edge" for a reason, it is your blood (money) that you lose. I see no reason to ever get something when it first comes out for this very reason. I let others take the hit and then when they get the bugs out, I will purchase. I'm done with being the quality control tester.
So do you think we will see new cards or a refresh next year as I am still with my 1080ti and I want a upgrade
@@maxsettingsnocompromises5724 you want to upgrade 1080ti? Nice
@@maxsettingsnocompromises5724 My guess first quarter/half of 2020. Im in same boat. Waiting for 3000 nvidia series with a 1080ti
From the look of the card, it seems you can solder another memory chip on it. Of course the bios wouldn't be happy, but there may be a fix... I kinda want to test it.
1:50 that what i talking about Speed Of Light (The Light)
First time I've heard someone pronounce BSOD like it's a word in 20 years. Great video though. :)
There's a good indie game called Primordia where it's pronounced in this fashion.
Because Linus keeps dropping them
Edit 1: when the likes hit 666 do not like this comment
Edit 2:YOU DISOBYED ME I GUESS 888 LIKES!?!
Linus would drop to his knees for any bloke
Which one? The clumsy or the anti Nvidia?
I tell my kids, “Don’t Linus that ice cream cone, because you’re not eating it off the ground.”
That edit is wishful thinking
Edit: When the likes hit 0 do not like this comment
Roger.
my rtx 2080 artefacts when overclocked to +905mhz on the memory with the auto core overock. starts showing blobs of bright red and green in random spots on the screen that grow into larger and larger blobs
hey man, I've put the card on water cooling, but when I play any game V-ram usage goes to the max... any help for this? thanks I did run heaven benchmark and did it with no problems at ultra rez... tried pretty much everything in that benchmark with a max temp of 53"
Failures are NOT only temperature related. Over the years there have been many cases where due to contamination, or new processes with smaller dimension designs, that over a period of time the slew rate of the chips drops. This means that the items will still operate well at LOWER clock rates, but not at the designed clock rates. The best thing to do is to look at the actual signals on the data lines with a O-Scope.
This was even seen back in the Atari 2600 days with some releases of various game cartridges. When new vendors were sought for the high volumes of ROMs needed for the games. Occasionally a new vendor would produce a chip that just did not perform long term as intended. With bench tests, it could be seen that the memory data outputs would not have a good stable output at the normal clock speeds the Atari 2600 ran at, often only reaching up to the "undefined" logic level voltages, before the next memory data addressed would be output to the data lines. Clocking the chip at a lower data rate revealed that the data was in fact still fully readable when the addresses were held longer to get stable normal logic levels on the data lines.
The same has been seen with various TTL and CMOS family logic chips with some production lots from various vendors, usually from new production lines. CPUs, Support chips, etc have also NOT been immune from these situations. So this is not a surprise to see in the world of fast evolving tech going into a very high volume and competitive world of Graphics cards.
"The best thing to do is to look at the actual signals on the data lines with a O-Scope. "
How would you do that? A decent oscilloscope ends at 100MHz..
@@stigthe5531 Currently, Tektonix makes scopes with a bandwidth up to 80 Ghz, and Lecroy makes them up to 100 Ghz. No they are NOT cheap, they are DSO scopes with real-time display capability.
@@denshi-oji494 that's just ridicules :0
my old oscilloscope can see max 30 Mhz....
Oscilloscope of that range would cost you like a car(or two)
Yes lots of money. It is hard to get real pricing on the very high end models, as they are typically ordered directly through the companies due to the price and limited market. but as a quick idea... expect to pay between $10,000 and $20,000 USD for the lowly 1 Ghz models (depending on specific features) 2 Ghz are about $25,000
And this is why I don't own "bleeding edge" hardware......
My 770 lasted pretty long for me and recently got a 1080 for my 120hz 1440p screen gotta say i think i dodged a bullet there bc i was seriously weighing my options. Same reason why im waiting til next generation for the new i5 to release.
i waited until 1080Ti was released before buying 1080. Sometimes bad shits happen.
There's nothing wrong with owning bleeding edge hardware. What's wrong is willing to own bleeding edge hardware that is known to fail a lot.
I wait a few months let others test it out
@@bgtubber yeah there is man its just like preordering games, if youve got the money and dont think returning things is a hassle, then sure. Otherwise you're buying hardware to be a guniea pig, might work out, might not. If you buy stuff thats slightly older, other people have done this testing for you... you know if average lifespan for card is under 1yr because a bunch have already died.
But yeah let me know how it goes dude maybe next year or in a few months if y'all are still stable I might consider new hardware. Otherwise the hassle isnt worth my time IMO
I just returned the nvidia rtx 2070 founder edition yesterday before my return policy expired.. worked for about 3 days. Ran it on 2 monitors. 2ndary mon wouldn't turn on or it would randomly turn off.. next I just used it on single monitor (Acer pred2 set at 120hrz) . Games started crashing to desktop, followed by crashing PC, to Bsod. ..which.. followed to failure to post windows..(I would have to reinstall windows completly and wipe PC partitions).2 weeks of trial an error (corrupt os, ram, even mother board) I finally took the card back to best buy and exchanged it for evga rtx 2070. With fresh install of windows and driver... no issues as of yet..
So if you do experience what I went through.. it's the GPU. Great video
Not only 2080 Tis. I have a 2070 and I also had that problem, but they switched it with a working one.
My 2070 artifacting after a week... horrible mistake buying this
EDIT: Got the card replaced and so far so good... fingers crossed
It just CRUSHED
2hours for me
Duncan Scholtz lmao😂
Do you happen to know off hand the equivalent 10x card performance for each 20x card? Like the 2080=1080(ti?) and the 2070 = 1070ti and the 2060 = 1070 non ti? Is that it ?
Fee I would say the 2070 is more like the normal 1080
He meant to say, that you can downclock the momory and keep using it, then one month before your warranty expires, RMA it and get a new series card.
or better yet, RMA it now. Why wait? you got a defective product, get it replaced, get it fixed.
@@navi1661 yeah that's why he bleeped it out
@@navi1661 because it's a 3year warranty card ! this means you will not recieve an RTX2080 TI as a replacement but it's successor ;)
@@sandidbouha3886 unless they screw you around and give you another rtx card.....that also.sucks.. because they'd probably be looking to dump them by that point.....
@@sandidbouha3886 Whoa! That's news to me! I've only ever got a replacement or a refund at best. How does that work? So they replace it with a new card with roughly the same value? I certainly wouldn't accept a much slower card.
7:34 is there supposed to be a memory chip there? Is it missing or come off when you opened it?
All 2080 Ti's are this way, 11G vram, so 11 gddr6 chips.
As a side note the comment on the video as to the amount of circuitry on the card is mainly due to it being a Gigabyte card, its common for them to do this, not all 2080 Tis are this busy.
Hi man, whats that onscreen information monitor you're using and everyone uses?? I can never find out!!
Ahh Micron, back in the day I got a 1070 when they were first out and it had Samsung memory, few weeks later I got another (1080ti wasn't out yet) to SLI them. The second card had Micron memory. It was unstable on stock, forget about any overclocking, and only after few weeks when a bios update was released it was finally able to be properly used. So I guess they didn't learn their lesson yet and are still going with the cheaper memory. I was ready to upgrade to 20 series but now I think I'll just wait. $1400 for 2080ti in US and I gotta downclock it to make it work? Yeah, no thx.
Samsung and Hynix make higher quality memory than Micron do
@@fayenotfaye just buy one because it JUST works right? Get real rude
@@fayenotfaye fuck that shit. If i buy a product i expect it to work out of the box. If i wanted to gamble on a 50/50 chance i would play the lottery.
@@DarkoPetreski If the lottery were a 50/50 chance I'd play it too! xD
@QuickNET Tech
50/50 gamble at a cost of 700$...
Doesn’t sound too bad.
On the plus side, if anyone needs a job, Micron has several QC positions opening up.
Although this time they might require previous experience...
So, any upates? was it due to micron memory failure? I had the same problem with RTX 2080. I am about to replace with Samsung Chip if it confirmed due to Micron failure
Is it on all brands that make the RTX models of GPU's or just specific brands.
I have had my MSI RTX2080 for 4 months now and has not missed a beat
Here
not yet
There are good ones, but there’s a decent amount of falty ones, for now your lucky
u wanna send me money m8?@Brad Viviviyal
Sadly I spent the last few days trying to figure out why games were insta crashing or loading in for less than a minute with artifacts and then crashing....it's a roughly two month old RTX 2070, and the memory downclock worked.... I'm just done
so does downclocking it fix the problem for you because i want to get a 2070 but im scared right now if i should buy it in case of stuff like this. but if there is a fix and that works i will buy it
@Cryotic. I have one myself that also had some isues and downclocking with afterburner works indeed, but if you are not sure don’t do it. You know, a 1080TI would be just as good except for rtx but you don’t need that
Emisminem ok thnx
Cryotic. Update, it only worked for 3 hours then my game crashed again..
Emisminem well 1080 it is
I had this same exact issue when I got the first batch of 980ti and EVGA exchanged it for me.
Really hope this gets fixed by May, I'm hoping to buy one around then.
I’m so glad I watched this video, thank you for making it and explaining it so objectively. I have a RTX 2080 that worked perfectly for 2 months, after I built the system. All of my other hardware should have made a perfect environment for this card to work at full potential. I have never OC this card yet, and temps were peaked at 53°C. I received artifacts and crashes all of a sudden during AC Odyssey (graphic detailed, yes, but shouldn’t have stressed this card or the system much). I reached out to Nvidia, Ubisoft, and Gigabyte (maker of my card) and couldn’t resolve the issue. I just sent it back for RMA, before finding this video. I kind of wish I still had it so I could have tested dropping the memory.
As I’ve seen elsewhere, this seems just to be a possible downside of rushing to the latest tech. I’m not at all upset with Gigabyte about this, as long as my card is fixed upon return. Yes, I believe that manufacturers should work this out before full-scale production, but at least for me I cut some small slack for them if the issue is addressed promptly.
Hey do u got a new card? And did you ever had the problem again?
@@Kidemeye Gotta love the no response back.
My 2080 kept hitting me with a BOSD. Luckily I was able to get a replacement. Hopefully this new one doesn’t continue.
Edconox Blue of Screen Death. Obviously.
what did the bsod say ? mine too is hitting me with critical_process_died
any news?
Did you still have problems with the new one?
It's not necessarily a thermal issue, not directly, and could well be a timing problem. Certain residues (flux, etc) left over from the manufacturing process can be relatively inert at normal temperatures, but start to flow under the thermal conditions of benchmarking, changing their dielectric properties. This can lead to increased capacitance between pins of the memory module, especially in a BGA package, changing the rise and fall times of high frequency signals.
mine just started malfunctioning, going through RMA hell right now
and i just did the same as you, turn down the clock speeds, turns out its the memory as well. couldnt run any benchmark before but after tweaks it works
They've sent me four replacement cards and all four have broken the same way in under a month
7:43 looks like alien Technology seems pretty advance lol
you are simply the best, only you could have cover this aswell
i have a RMA PNY 2060 super XLR8 i had those texture flicking shimmer and shadow and lighting problem on a brand new PC,
try different board and cable monitor and i also did use a second monitor and graphic card
probably bcuz the speed of the vram on them were too overclock, i dunno, but im glad i send it back this was micron memory too.
so i ship it back in USA waiting for a replacement, PNY guys were comprehensive.
i just wish this time i get a working one with different memory but i doubt....
now i don't know if this happen few month later if this still going to be as easy to be replace
most of the ppl still keep them bcuz they think it's a drivers issue....
Is there any way you could give some advice on testing the GDDR6 memory itself? Are there any software diagnostics that can be run to test just that component directly? I understand the basic test of running down the Memory Clock using Afterburner and seeing if it affects benchmark stability, but that feels like a very crude test and I'm not entirely sure what it's actually telling us. Moreso for me, because my problems are not consistently reproduceable. I've been running Furmark, Cinebench and Octanebench for testing and probably only 30-40% of the time any of the 3 will trigger a total crash (either BSOD / artefacting / total system failure and reset). So I don't think I can really get a reliable indication by reproducing that test here. Is there an equivalent of MemTest64 that can be run on a GPU's onboard memory that you know of? Any advice would be appreciated. I will still RMA this card but I'd really like to have understood the nature of the problem, especially as there are many cases online of people that have had similar problems after 2 or 3 replacement 2080 Ti's. I really don't want to deal with this problem again, I've spent 2 whole weeks trying to diagnose and fix it, or at least rule out software as the issue. Having tried pretty much everything possible, I'm certain it's a hardware fault. But I'd love to know what exactly the problem is!
Maybe so, maybe so - I'd still return it so Gigabyte's warranty people can investigate and determine the exact fault and kick in gear any corrective action for future production runs. Maybe I'm too much of an ideologist, but I trust a brand like Gigabyte would do this as opposed to simply replacing the card blindly (and any reputable company should be doing this - I was a warranty analyst for a big company that actually did do this). :-)
@Reggie Reginato
At least yours aren’t ASUS
I’m sure Gigabyte know all about the issue already from the faulty cards that they have had returned before this one failed. No company is going to tell the public what the problem is in detail. They have an RMA procedure and that’s where it begins and ends for customers that have faulty cards. The discussions with the component manufacturers behind the scenes would be interesting though.
@@moebius2k103 they can't rely on form posts for evidence - too high of a risk for tainted and incorrect information. And your right, companies won't discuss problems of this nature publicly for a large number of reasons (competitive through public mass hysteria). Even if it's a problem that is shared by competitors using the same component manufacturer for the same components.
Reggie Reginato ya, I agree, even if he doesn't feel the need to stir the pot and ask for a replacement, he should at least send his contact a message saying what happened and maybe link to the video, they may want to get it back to look at it if they are still trying to figure out the issue. They may also want him having a working card for benchmarking and testing since thats why they gave him a review unit to begin with. Doesn't hurt to send a matter of fact, this is what is happened and this is how I have gotten it to stop, message just for another input into the issue.
@@FireMrshlBill They've had hundreds if not thousands of these returned already, they know what the issue is, returning his one won't give them any info they don't already have. This is an OLD issue now, replacement cards are being sent out with samsung memory and have been holding up fine so far.
I'l just wait for a new generation of gravis cards. Yup.
Looks like a massive populated city when you opened it up
Yet probably works alot more efficiently than nearly any Cities Skylines city built to this very day.
Found this video, good one and i want to ask question. I found out my RTX 2070 super doing lighting artifacts especially on snow and also there are like purple and blue colors around it however i noticed it is related only to one game. I did benchmark testing and other games and there i didnt see anything similar to occur, could this issue be related only to that one particular game or not? thank you.
Crazy this is still an issue. Even hearing the issues on some of the new "super" cards as well. My Zotac AMP 2080Ti (micron) has been rock solid for over a year now. With my luck, I can't believe I dodged that bullet.
Edit: you don't have to tear down the card to find out whos VRAM it is, just run GPU-Z and Samsung modules were seeing almost as many failures as the micron Vram.
Silicon Lotto with expensive cards
Is this still going on?
My 1080 strix (2 years warranty left) the RGB only lights up half the card and the White is pink. Could I RMA in a year when there's no stock and get a new one.
"why you little....!!"
@@lolbots Oh the Simpsons. DOH!
I'm guessing they would replace it with a 2070.
Yes. They will also probably give you a newer one.
@@MG-ks1qg I didn't even notice the RGB wasn't working properly until I vertically mounted it (8 months soo pissed off). stupid me I should have tested it outside of the case soon as I bought it but I guess it's going to work out better in the long run if I get a new card.
What are the pc specs u used with this test? Cpu/ram/psu?
So when you downclock the memory, only the minimum framerates goes down? I've been working my 1080ti that I just put an Accellero III cooler on, messing with OC settings. Would be nice to know exactly how memory clocks works with compute clocks...
Well, my wait for AMD Radeon VII seems reasonable, mostly for content creation and longevity.
Yup. I can't wait for them. I think it would be my first AMD card?
Is there any way of finding out the memory brand before purchase? Or any info on which manufacturer uses micron/samsung memory?
When you receive it, test it on GPU-Z, it'll tell you what the memory is. Just return it as 'I changed my mind' if it says Micron... Lol
Lol indeed!
@@Addy187a1 You could also overclock the snot out of it as surely if it was part of the bad batch it would fail quickly. I doubt ALL Micron memory is bad.
@@Addy187a1 Thats a lot of returning lol
Hey i have a question @Tech Yes City, Im the owner of one RTX 2080 Super, and i got trouble it Artifacts and Render objects fail in games, also my RTX did crash my pc several times seconds after start launch some game, just the time too open the game and "Voila" BSOD or some times, shuting down my system, staying in loop for several minutes and rebooting too bios.
I send it too the manufactuer but, i didnt receive any note about what cause that failures :/
In adicional notes, i can say that i use my GPU in base clocks, and i do not do OC on my system, i also pay an extra on a expensive PSU for get my system proctect but even it all this cares, this is my 3rd GPU in less of 3 years that get artifacts failures, do you have any idea of, what causes gpu fails/die?
Thank you Sir
Very cool video I learned a lot!!