Got a Bad 2080 Sent it Back & the New One Has A Fan Problem....

[Sithtiger]

So I had this thread first if you want to look about my previous 2080. So I received a new MSI GeForce RTX 2080 Ventus 8G and it works beautifully and was working fine until I upgraded to the 418.91 drivers and played Apex Legends. After playing for around 30 minutes or so, I had a hard reboot. I knew it was the card, because the last one did that all the time. The difference is it had been fine until the new driver and Apex Legends. I'm not sure if it's the driver or Apex Legends.

Now, I did solve the problem by simply starting MSI Afterburner, but I shouldn't have to do that and I wasn't even O/Cing either. Should I send this one back too or just stick with it. What's the deal with the fans? I could pop my MSI GTX 1070 in there and run it at stock and it would be fine. I'm wondering if a driver is doing this. Also, MS just did an update, so it's one of those 3 things. I guess I need to try another game first.....sigh.

Do you guys agree? I know I need more testing, but it shouldn't have done that. Also, has anyone heard of this happening with Apex Legends or the new drivers causing spontaneous reboots are RTX 2080's overheating or fan problems? Any of it...

 

[Johnx64]

Rollback to the previous driver.

 [Apex Legends]: DXGI_ERROR_DEVICE_HUNG error message may appear and then the game crashes.[2503360]

 

[Sithtiger]

The 418.91 Nvidia drivers caused my GPU to stay in a perpetual heavy load condition. It wasn't O/Ced either.

You can tell this by downloading GPU-Z or just by opening Afterburner or Precision. Any of them. If you're not doing anything, you'll see your GPU is on full load as if you were playing a game or something else that was graphic intensive. It never fluctuated. I don't know if anyone else has noticed this, but it would be easy to miss as the fans only ran at around 23% the whole time. The fans never sped up to lower the heat and that's why it crashed the core and memory did. I played Apex Legends (the game that crashed it) for over 40 minutes. My 2080 never overheated. So, if anyone has a crashing problem with their GPU's all of a sudden, check to see if it's running at normal load speeds at idle. If it is, then go back to 418.81 and that should fix the problem. I should add that I was mirroring my monitor, but I even disabled and unplugged the HDMI to my TV. So, even that didn't work.....that is for allowing the card not to run under full load. So this happened to a desktop PC too. I'm sure some people don't even know their card is running at full load unless they play a game and it crashes. The difference between 418.91 and 418.81 is when playing Apex Legends or another game, your GPU will run at lower speeds (core, memory and voltage) if it doesn't need all the power unless of course you set it to do so, but I wouldn't recommend that as it will just lower the lifespan of your GPU even at stock speeds if it's running at full load all the time.

 

[Sithtiger]

**UPDATE**

I have to submit a correction and an apology to Nvidia. The situation I was having was not caused by the 418.91 drivers, but rather the problem was.....sigh...ME being an idiot. In Nvidia Control Panel in 'Manage 3D Settings' in the "Power Management Mode" I had inadvertently changed the setting to 'Prefer Maximum Performance' instead of leaving it to default or set it to Adaptive. I had thought I had restored all settings to default, but apparently, I did not, so to reiterate. There is no problem with the 418.91 drivers running the GPU while idle. I had simply set it to Prefer Maximum Performance, which would cause the GPU to heat up and could cause games to crash ever at default speeds. Note, I have not tested this yet, other than verifying that the GPU is, in fact, running at normal speeds at idle, so please disregard my stupidity in this.

I won't make a stupid mistake again regarding that. Perhaps a different one, but not that one. Has anyone seen "The Davinci Code"? I'm punishing myself as Silas did with a Cilice belt around my leg, except mine is made with older GPU's. I have two 970's and my1070 in the belt, so I shall never make that mistake again. I haven't checked this in games yet, I'm about to, but I'm very confident that everything will be OK now. If for some reason this occurs again, I will report back here, but unless I do, consider the matter closed....on my part anyway. There still is the matter of the Cloned Display causing similar-like issues, but this is a different problem.