Do YOU have Problems with YOUR RTX card?

[DooKey]

I'm actually done stress testing this card. I've played a ton of VR games using my Vive, and so far my PC has seen a lot of gaming activity since the day I purchased it. I have a hard time believing that 100% of the RTX2080ti cards are faulty, but I guess only time will tell.

They aren't all bad or going to go bad. This generation seems to have an anecdotal higher failure rate than previous high-end cards.

 

[Slade]

All products have failure rates, it just seems that the 2080ti has a higher than expected one. Let's be honest about GPU technology that it is getting pretty big and complex. A GPU in the modern age pulling 350W of power is insane yet here we are.

 

[FizzBeauc]

GIGABYTE RTX 2080 GAMING OC on it's way, should get it installed tomorrow evening and be able to do some stress testing. Fingers crossed!

 

[cybereality]

Good luck. Hope you enjoy.

 

[VF31Tomcatters]

Just bought an EVGA 2080 Ti Ultra

You guys are making me nervous

 

[FizzBeauc]

Just bought an EVGA 2080 Ti Ultra

You guys are making me nervous

Install card, cross fingers, play games.

 

[Slade]

Just use it and don't stress too much over it. I've only had 1 card failure in 18 years and was mem related on it. Others have seen multiple in 1 generation. That's what warranties are for.

 

[Dacvulcan]

I ordered my RTX 2080 Ti directly from NVidia a month ago and was surprised how quick it came. I am running it on an Asus TUF Z390M board with a i7 8700K overclocked to 4.8GHz. I have turned the system on every evening for the last three weeks and have run benchmarks on the card - Heaven 4.0 etc. I am driving a 28" 4K monitor at 3840 x 2160. I downloaded the EVGA X1 app and can watch the temps during the benchmarks. The RTX has not exceed 75 degrees C and the card's fan haven't exceeded 55%.

I have ordered an Alphacool Eiswolf 240 GPX Pro water cooler. I don't plan on overclocking the card but just want the additional insurance plus I can add more LED fans to look at in my Thermaltake Core P3 case.

I replaced two EVGA GTX 1080 FTW Hybrids with this card and am totally satisfied with the performance.

 

[Legendary Gamer]

So.... I picked up an EVGA Black Ed 2080Ti, it's running Micron memory.

I'm not at home to post tech info atm, fyi.

I was playing Battletech at 4K 4:4:4 Chroma via HDMI on my old 1080Ti waiting for the new card to show up and it exited the game to desktop repeatedly (possibly dying). Didn't have that issue on the 2080Ti when it arrived.... However, I see odd glitches on the screen running 4:4:4 Chroma on the 2080Ti that I wasn't seeing on my 1080.

I'm using a known good, week old 18gbps cable (10ft) ... Might have to move the tower and use a 6 footer, as I know distance is a factor with 4K over HDMI.

The glitches are fast, like a partial second screen flash a couple times a minute. They diminished (but still happens) when I dropped to 4:2:2 but it just chaps my ass that this didn't happen with my 1080Ti at all and it does with my brand new 2080Ti.

It's also definitely heating my room even at stock settings. I attempted to shave 5-10% off the power allowance with the EVGA utility but that introduced micro stutter into the game (kinda figured at lower power the card would clock lower and run cooler)... So I kicked it back up to stock.

Anyone else seeing intermittent screen flashes at high Chroma settings?

 

[brentsg]

I was using my 2080 Ti and I had one incident where everything seemed perfect, when done gaming I shut down normally. But then the next boot, I got the "something is wrong" shit. Unfortunately I hit the power button and walked away for a bit so I just got a glimpse of it when walking back in, thus I didn't get to read the specific message.

The other weird one was that I simply lost video out once. Playing a game and boom, gone. I tried to change display inputs and such, as the PC was still running. Eventually I had to do the hard shutdown and reboot.

Now I don't trust it.

 

[Legendary Gamer]

The other oddity I forgot to mention was that the GeForce Experience no longer functions. No matter how many times I reinstall it, it keeps kicking a "Something went wrong" on load...

It's dead Jim... My God... What will I do without automatic configuration!?!

Lol

Oldskool manual adjustment time

 

[Legendary Gamer]

I was using my 2080 Ti and I had one incident where everything seemed perfect, when done gaming I shut down normally. But then the next boot, I got the "something is wrong" shit. Unfortunately I hit the power button and walked away for a bit so I just got a glimpse of it when walking back in, thus I didn't get to read the specific message.

The other weird one was that I simply lost video out once. Playing a game and boom, gone. I tried to change display inputs and such, as the PC was still running. Eventually I had to do the hard shutdown and reboot.

Now I don't trust it.

I don't trust it at all... I think I spend more time looking at my open PC chassis half expecting the Fu#!5&$ thing to light the surrounding components ablaze...

 

[brentsg]

So.... I picked up an EVGA Black Ed 2080Ti, it's running Micron memory.

I'm not at home to post tech info atm, fyi.

I was playing Battletech at 4K 4:4:4 Chroma via HDMI on my old 1080Ti waiting for the new card to show up and it exited the game to desktop repeatedly (possibly dying). Didn't have that issue on the 2080Ti when it arrived.... However, I see odd glitches on the screen running 4:4:4 Chroma on the 2080Ti that I wasn't seeing on my 1080.

I'm using a known good, week old 18gbps cable (10ft) ... Might have to move the tower and use a 6 footer, as I know distance is a factor with 4K over HDMI.

The glitches are fast, like a partial second screen flash a couple times a minute. They diminished (but still happens) when I dropped to 4:2:2 but it just chaps my ass that this didn't happen with my 1080Ti at all and it does with my brand new 2080Ti.

It's also definitely heating my room even at stock settings. I attempted to shave 5-10% off the power allowance with the EVGA utility but that introduced micro stutter into the game (kinda figured at lower power the card would clock lower and run cooler)... So I kicked it back up to stock.

Anyone else seeing intermittent screen flashes at high Chroma settings?

I’d generally expect the room heating to be a wash with the 1080Ti.

 

[cybereality]

The glitches are fast, like a partial second screen flash a couple times a minute.

I was getting this too on my 4K TV setup. The screen would flash black at random times. I thought it was the cable, and I tried like 4 different high-end cables and the problem persisted.

It ended up being some sort of interference. I have like 1,000 cables all under the desk, and they were all touching (including power cables) and I think this was the issue.

I used twist ties to separate the main cables for my TV, so they were tied to the desk legs at different points so they were not touching. Then the problem went away.

 

[Legendary Gamer]

I was getting this too on my 4K TV setup. The screen would flash black at random times. I thought it was the cable, and I tried like 4 different high-end cables and the problem persisted.

It ended up being some sort of interference. I have like 1,000 cables all under the desk, and they were all touching (including power cables) and I think this was the issue.

I used twist ties to separate the main cables for my TV, so they were tied to the desk legs at different points so they were not touching. Then the problem went away.

I don't have the same level of wiring spaghetti, however, you make a good point. I changed my cable placement last night and the issue went away at 4:2:2. Have yet to try it back at 4:4:4 Chroma. I'm wondering if my Klipsch bookshelf speakers might be causing an issue . Worth a look.ThankThanks thanks for the idea !

 

[Legendary Gamer]

I’d generally expect the room heating to be a wash with the 1080Ti.

I actually did more testing with underclocking the 2080Ti last night and the heat issues totally disappeared. I'm theory crafting / reaching here but what if the main issue with the RTX cards is that Nvidia TOO aggressively paced the boost clocks and the chips and boards themselves cannot actually handle the load? Likely done because at stable levels of performance the card is only 10-15% faster than a 1080Ti . I dropped my power about 15% last night and placed the target thermals at 74-75c and the card pretty much never got that even under load. Gonna have to invest in an infra red thermometer to do some real testing.

 

[Mchart]

I actually did more testing with underclocking the 2080Ti last night and the heat issues totally disappeared. I'm theory crafting / reaching here but what if the main issue with the RTX cards is that Nvidia TOO aggressively paced the boost clocks and the chips and boards themselves cannot actually handle the load? Likely done because at stable levels of performance the card is only 10-15% faster than a 1080Ti . I dropped my power about 15% last night and placed the target thermals at 74-75c and the card pretty much never got that even under load. Gonna have to invest in an infra red thermometer to do some real testing.

I sit at like 70c full load at 100% power target with a linear fan curve. I don’t think heat is much issue here.

 

[Hagrid]

I sit at like 70c full load at 100% power target with a linear fan curve. I don’t think heat is much issue here.

What are your other gpu's components temps like?(Memory, VRM, etc)

 

[Starrbuck]

I sit at like 70c full load at 100% power target with a linear fan curve. I don’t think heat is much issue here.

I agree. I'm at 54-55C at full load under water with 112% power.

 

[Legendary Gamer]

I sit at like 70c full load at 100% power target with a linear fan curve. I don’t think heat is much issue here.

It definately seemed like an issue on the black edition. I was making me sweat the level of heat that it was generating in a 10x10 room.

Might depend on the card and the cooling situation. Also guessing the black Ed is a lower binned chip due to it's price being 999.

 

[Legendary Gamer]

I agree. I'm at 54-55C at full load under water with 112% power.

This is what I'm talking about. There is no way to make a comparison here when you're rocking water cooling and I'm just using the stock, HSF that they throw on both the 2080 and 2080Ti cards AFAIK.

I want to see what the stock card does without water or an enhanced HSF. My card was listing thermal targets for 82c under 100% power load and honestly.... The thought of pushing the card never entered into my mind after I saw the heat spike at stock (non FE) clocks. It can only mean that the FE clocks fucking cook the electronics on these cards...

 

[Furious_Styles]

This is what I'm talking about. There is no way to make a comparison here when you're rocking water cooling and I'm just using the stock, HSF that they throw on both the 2080 and 2080Ti cards AFAIK.

I want to see what the stock card does without water or an enhanced HSF. My card was listing thermal targets for 82c under 100% power load and honestly.... The thought of pushing the card never entered into my mind after I saw the heat spike at stock (non FE) clocks. It can only mean that the FE clocks fucking cook the electronics on these cards...

If that's true we might see a stealth cut in boosting voltage/clocks in a forthcoming driver release.

 

[brentsg]

This is what I'm talking about. There is no way to make a comparison here when you're rocking water cooling and I'm just using the stock, HSF that they throw on both the 2080 and 2080Ti cards AFAIK.

I want to see what the stock card does without water or an enhanced HSF. My card was listing thermal targets for 82c under 100% power load and honestly.... The thought of pushing the card never entered into my mind after I saw the heat spike at stock (non FE) clocks. It can only mean that the FE clocks fucking cook the electronics on these cards...

I don't follow the logic. The FE previous cards have been boosting to 82C/84C and just sitting there, adjusting clocks to keep the temp steady. Nothing has changed here except the cooler format.

 

[Legendary Gamer]

I don't follow the logic. The FE previous cards have been boosting to 82C/84C and just sitting there, adjusting clocks to keep the temp steady. Nothing has changed here except the cooler format.

I have a bullshit HSF on my card is what I'm referring to.

 

[Mchart]

What are your other gpu's components temps like?(Memory, VRM, etc)

No clue. I have no way of telling. All I know is that I did build a new computer when I got my 2080ti because my prior computer case/build was really designed for a blower style GPU cooler - Not the way the 2080ti cooler is. My temps were WAY higher with my prior build. In my new build with a fractal design R6 case there is enough air coming in front blowing straight into/over the GPU that it's operating basically 10c cooler.

I'm sure in cases without proper airflow setup the card is overheating in certain spots.

BTW the temps I'm saying are just the standard FE cooler.

With my current thermals I can move the power slider to 120% and I basically sit at 80c at full load.

 

[Hagrid]

No clue. I have no way of telling. All I know is that I did build a new computer when I got my 2080ti because my prior computer case/build was really designed for a blower style GPU cooler - Not the way the 2080ti cooler is. My temps were WAY higher with my prior build. In my new build with a fractal design R6 case there is enough air coming in front blowing straight into/over the GPU that it's operating basically 10c cooler.

I'm sure in cases without proper airflow setup the card is overheating in certain spots.

BTW the temps I'm saying are just the standard FE cooler.

With my current thermals I can move the power slider to 120% and I basically sit at 80c at full load.

That could be the problem. People are looking at 1 temp and it says fine, so no prob. But there are many components on the board that could die due to excessive heat.
I think GPU-z has a few temp sensors? Not sure how they work with the RTX series.

 

[Slade]

I think if heat alone was the cause, pretty much every 2080ti that hit high thermals would be dying. There's more to this than just bad memory and thermals that are causing early deaths of cards. I'm into week 4 of EVGA 2080ti and avg 70's temp stock with doing a 130% power hitting 82C under full load. It's been on 24/7 since I picked it up.

 

[Mylex]

Shut it down I dare you lol. Just kidding I think the issue with these card are multiple things but if you stay away from a reference card you have a better chance at finding a good one.

 

[Legendary Gamer]

I think if heat alone was the cause, pretty much every 2080ti that hit high thermals would be dying. There's more to this than just bad memory and thermals that are causing early deaths of cards. I'm into week 4 of EVGA 2080ti and avg 70's temp stock with doing a 130% power hitting 82C under full load. It's been on 24/7 since I picked it up.

My experience with high tech components is that Heat is the leading cause of the death of components, period. Now, a die shrink can do one of two things 14nm to 12nm ... It can allow you to save power or clock higher (depending upon the refinement of the node). A good example of this is how we see TSMCs 14 to 7nm process and what it can deliver. There are a bunch of articles on that out there .

The way I see it is that Nvidia went cheap, the kicked out a card with some questionable tech on it (raytracing is cool but.. the penalty on first gen hardware is brutal). It's not like stepping from a 2D game to a 3DFX Voodoo 1 that kept my game at a solid 30+ fps while making everything look awesome. It's taking an awesome game and adding something that cripples your performance... Anyway I'm getting way off topic here.

So save power or clock higher are your choices for a die shrink... Nvidia is not saving power, in fact they're cranking power higher and pushing the clock at the same time. In my mind that violates the advantage of a die shrink and seems plausible that they're just pushing these damn cards too hard.

These cards on 7nm will be sweet. Now... I don't think 12nm is the right node for these cards to sit on. So, you have the die cranking out heat, the dram hitting max thermals and everything on the damn board is pretty much on fire.

Heat kills electronics. Pure and simple regardless of what the specs "say it's rated for" What I'm curious about is how many of these 2080Ti's under water have failed?

 

[Dayaks]

My experience with high tech components is that Heat is the leading cause of the death of components, period. Now, a die shrink can do one of two things 14nm to 12nm ... It can allow you to save power or clock higher (depending upon the refinement of the node). A good example of this is how we see TSMCs 14 to 7nm process and what it can deliver. There are a bunch of articles on that out there .

The way I see it is that Nvidia went cheap, the kicked out a card with some questionable tech on it (raytracing is cool but.. the penalty on first gen hardware is brutal). It's not like stepping from a 2D game to a 3DFX Voodoo 1 that kept my game at a solid 30+ fps while making everything look awesome. It's taking an awesome game and adding something that cripples your performance... Anyway I'm getting way off topic here.

So save power or clock higher are your choices for a die shrink... Nvidia is not saving power, in fact they're cranking power higher and pushing the clock at the same time. In my mind that violates the advantage of a die shrink and seems plausible that they're just pushing these damn cards too hard.

These cards on 7nm will be sweet. Now... I don't think 12nm is the right node for these cards to sit on. So, you have the die cranking out heat, the dram hitting max thermals and everything on the damn board is pretty much on fire.

Heat kills electronics. Pure and simple regardless of what the specs "say it's rated for" What I'm curious about is how many of these 2080Ti's under water have failed?

2080ti has the same stock TDP as the 1080ti. These things are dying at stock. And relatively the temps aren’t that high, same max temps since forever and are much lower than some of AMD’s cards.

I think the issues are less simple.

 

[brentsg]

2080ti has the same stock TDP as the 1080ti. These things are dying at stock. And relatively the temps aren’t that high, same max temps since forever and are much lower than some of AMD’s cards.

I think the issues are less simple.

This is the right answer. The TDP and temps that we're talking about have been commonplace for several generations now.

 

[cybereality]

I actually did more testing with underclocking the 2080Ti last night and the heat issues totally disappeared.

What does your fan profile look like? The default one is trash and you need to make your own.

This is the one I'm using on a 2080 Ti FE overclocked (+160 +500).

Keeps things pretty cool, on average around 75C while gaming. Since I'm on SLI, I have seen the top card get to around 85C in certain situations, but that is okay for short periods.

 

[Legendary Gamer]

What does your fan profile look like? The default one is trash and you need to make your own.

This is the one I'm using on a 2080 Ti FE overclocked (+160 +500).

View attachment 124133

Keeps things pretty cool, on average around 75C while gaming. Since I'm on SLI, I have seen the top card get to around 85C in certain situations, but that is okay for short periods.

Sorry I went off on a rant guys. I have my fan profile set to Aggressive and it looks exactly like the posted image.

I suppose part of me is just paranoid it's going to light my system on fire.

It's a helluva lot smoother experience playing 4k games tho. Finally will put the old 1080Ti in the front room should this thing last by some miracle.

You're right about the temps, they've been this way for a very long time.

 

[Bop]

Just chiming in with my experience. My RTX 2080 Ti keeps crashing to desktop in Hitman 2 on a specific level (Santa Fortuna) with an nvlddmkm stopped responding error in the Windows event log. Underclocking the GPU stops the crashing.

 

[N4CR]

Just chiming in with my experience. My RTX 2080 Ti keeps crashing to desktop in Hitman 2 on a specific level (Santa Fortuna) with an nvlddmkm stopped responding error in the Windows event log. Underclocking the GPU stops the crashing.

Almost like the P3 1GHz all over again...

Heard of a theory that power cycling/inrush may be causing some of the issue, many cards reported issues after boot.

 

[KickAssCop]

Just chiming in with my experience. My RTX 2080 Ti keeps crashing to desktop in Hitman 2 on a specific level (Santa Fortuna) with an nvlddmkm stopped responding error in the Windows event log. Underclocking the GPU stops the crashing.

Seems more like a driver issue if this is the only game it is crashing in.

 

[Mazzspeed]

Wow, these statistics based on the size of the sample group within [H] aren't good! I'm sort of glad I didn't spend all those $$ on an RTX card.

 

[Mchart]

Just chiming in with my experience. My RTX 2080 Ti keeps crashing to desktop in Hitman 2 on a specific level (Santa Fortuna) with an nvlddmkm stopped responding error in the Windows event log. Underclocking the GPU stops the crashing.

This isn't your card - The game is crashing for anyone on Turing. It's a driver and/or game conflict/issue. Unfortunately it's not patched yet.

The crash is very random, and underclocking may seem to help, but trust me it doesn't.

 

[FizzBeauc]

I decided to preemptively return the RTX 2080 and get a 2070 OC instead. Saved some $$ and potential instabilities. Either way, it's a far cry from my 1050 Ti!

 

[Legendary Gamer]

This isn't your card - The game is crashing for anyone on Turing. It's a driver and/or game conflict/issue. Unfortunately it's not patched yet.

The crash is very random, and underclocking may seem to help, but trust me it doesn't.

The drivers released the other day made a big difference. Also, the white screen tearing I was suffering at 4:4:4 Chroma disappeared. It looked like issues related to cable bandwidth and I had replaced the cable and dialed back to 4:2:2 to defeat it, now, finally, I enjoy it in all it's glory.

Less game crashes to the point I was able to game for several hours in an enjoyable fashion. Granted I was just hammering a heavily modified battletech session but it was stable with only one early CTD in 5-6 hours of game play. Not one crash every 45 min. Also helps Paradox/HBS patched their shit game stability as well.