Need help with a GTX 1060 not working properly.

[Squid04]

At first I thought my GPU had died, however I can see it in windows device manager, afterburner and nvidia controlpanel. 

I have done a clean install of the newest drivers with the help of DDU and checked by connections. Cmos has been cleared. 

• MSI GTX 1060 6G Single fan. 
• Only outputs with  DVI. No displayport. 
• Fans don't spin up. 
• Seems to be drawing power as it was hot. 
• Problem started after the PC had gone into sleep mode and screen never went on again. (Was using DP) 
  • Had similar issues for about a month, but it always started working again. 

[tarfeef101]

Oookayy.. this should be fun

 

1. Is it working in 3D applications at all right now (over DVI), or no (just like, the desktop and chrome and stuff)
2. Is it hot only under load, or even at idle?
3. Do the fans never spin, or just not spinning because it's not under load?
   1. What does the fan RPM say in afterburner? have you tried setting a fan curve to make them always on and seen if that works?

 

I would try to get as much info while using another GPU/integrated graphics first, then start using the GPU in question as needed to test it out. just in case. 

perhaps also try running a CUDA test while connected to integrated graphics, that'll help see if drivers installed properly and if the issue is strictly a display one

[Squid04]

15 minutes ago, tarfeef101 said:

Oookayy.. this should be fun

 

1. Is it working in 3D applications at all right now (over DVI), or no (just like, the desktop and chrome and stuff)
2.  Is it hot only under load, or even at idle?
3. Do the fans never spin, or just not spinning because it's not under load?
   1. What does the fan RPM say in afterburner? have you tried setting a fan curve to make them always on and seen if that works?

 

I would try to get as much info while using another GPU/integrated graphics first, then start using the GPU in question as needed to test it out. just in case. 

perhaps also try running a CUDA test while connected to integrated graphics, that'll help see if drivers installed properly and if the issue is strictly a display one

1. Working on redownloading Heaven benchmark

2.It hangs around 38 Celsius and 44 Celsius. 

3. The fan never spin, even though I tried to force it to hit max speed. 

 

I'm grateful for your tips!

Edited February 12, 2019 by Squid04

[Squid04]

41 minutes ago, tarfeef101 said:

Oookayy.. this should be fun

 

1. Is it working in 3D applications at all right now (over DVI), or no (just like, the desktop and chrome and stuff)
2. Is it hot only under load, or even at idle?
3. Do the fans never spin, or just not spinning because it's not under load?
   1. What does the fan RPM say in afterburner? have you tried setting a fan curve to make them always on and seen if that works?

 

I would try to get as much info while using another GPU/integrated graphics first, then start using the GPU in question as needed to test it out. just in case. 

perhaps also try running a CUDA test while connected to integrated graphics, that'll help see if drivers installed properly and if the issue is strictly a display one

So I ran the benchmark and it's saying it's running of the GPU. I also held my hand above it and it was radiating heat, so has to be working.
Weird issue is that the Displayport does not work nor do the fans spin...

[tarfeef101]

16 hours ago, Squid04 said:

- snip -

Okay, couple things:

 

1. I suggest you don't run anything intensive on that card right now. That GPU is not supposed to get that hot. 
2. As for fixing it, I see 2 good options
   1. RMA the thing. Easiest solution, and it clearly has problems that aren't your fault
   2. You could try flashing the BIOS, i doubt the hardware died. It is POSSIBLE electrostatic shock or wear on the connector killed the DP, and it is possible the fan connector got disconnected or they got ESD damaged as well. but not likely. Not sure how that will affect your warranty, so the safest option it just to RMA. But if this doesn't affect warranty for you, you don't care, or the warranty period is over anyways, that is an option that may help. (also checking to ensure the fan cable is plugged in for the GPU fans isn't gonna hurt, either. if you're comfortable taking the card apart, then it's a free excuse to replace the stock thermal paste with better stuff anyways

[Squid04]

8 minutes ago, tarfeef101 said:

Okay, couple things:

 

1. I suggest you don't run anything intensive on that card right now. That GPU is not supposed to get that hot. 
2. As for fixing it, I see 2 good options
   1. RMA the thing. Easiest solution, and it clearly has problems that aren't your fault
   2. You could try flashing the BIOS, i doubt the hardware died. It is POSSIBLE electrostatic shock or wear on the connector killed the DP, and it is possible the fan connector got disconnected or they got ESD damaged as well. but not likely. Not sure how that will affect your warranty, so the safest option it just to RMA. But if this doesn't affect warranty for you, you don't care, or the warranty period is over anyways, that is an option that may help. (also checking to ensure the fan cable is plugged in for the GPU fans isn't gonna hurt, either. if you're comfortable taking the card apart, then it's a free excuse to replace the stock thermal paste with better stuff anyways

Yeah, I stopped using it for intensive stuff. Don't want to dmg it further. I contacted the company I got it from and they were willing to RMA it. Unsure if it's covered by the warranty or if it's still there. But we'll see. It failed by itself happily and hopefully. I assume they want to run tests on it before hand, kind of hope the issue is unfixable as that would mean they'd have to hand over a new one, rather than repair this unreliable one. 

 

Very grateful for your help dude!

[tarfeef101]

2 minutes ago, Squid04 said:

Yeah, I stopped using it for intensive stuff. Don't want to dmg it further. I contacted the company I got it from and they were willing to RMA it. Unsure if it's covered by the warranty or if it's still there. But we'll see. It failed by itself happily and hopefully. I assume they want to run tests on it before hand, kind of hope the issue is unfixable as that would mean they'd have to hand over a new one, rather than repair this unreliable one. 

 

Very grateful for your help dude!

glad you got a solution

my pleasure. (feel free to choose a best answer, or leave informative reacts to posts that help you to show people your appreciation  )