Graphic Cards GPU or PSU, which is failing

mk76

Adept
GPU - MX RX580 Armor OC ( Out of warranty)
PSU - Cooler master V750
Others - Z170 Sabertooth + 6700K + GSKILL 2x16B @ 3000


Couple of months ago, my system crashed once with Asus surge protection message. That time I cleaned up the grill, fan, gpu etc and it ran fine thereafter.

Yesterday I faced following issues.

Observation 1
- While playing Far Cry 4 ( set at 2560x1080 @ Ultra ) > screen turned blank > after few sec POST beep > Asus surge protection triggered
- So I cleaned again, and it ran fine for few hours

Observation 2
- After couple of hours post cleaning, system crashed ( turn blank screen > few sec > restart ) every-time
--- Game would start.
--- Upon Windows load > login screen > press enter to focus pin window. However this behavior was not consistent
--- On running fur mark

=== I removed the gpu, and ran igfx ... all well
=== Removed AMD drivers (safe mode) , and reinstalled

Observation 3
- Post drivers reinstall, I ran Furmark at lowest res ( 400x300 .. i guess ). The window showed up for a sec and then the crash cycle triggered.

- I ran CPU benchmarks to generate load on PSU , it ran fine.
- Don't have a second system to test

Any suggestions what could be the issue? Is the GPU dead?
 
Me and few members here had issues with Farcry and other games at start of this year and windows updates were the culprit. I uninstalled all crappy updates and since then havent updated my win 10 and games running butter smooth.
 
1] is Earthing connection is ok with/at your place
2] did you check motherboard for any abnormality... is any capacitors change it's shape... my last 2 system (BOTH ORIGINAL INTEL Motherboard) turn bad and I found capacitors (plastic one), start swelling
3] check RAM
 
Here's what all I checked

With GPU
- Cleaned and applied new thermal paste to both CPU and Rx580
- Issue remains
- Upon boot, when login screen appears, I just pressed enter to focus on input field ... reboot. It takes 5 sec then post cycle begins
- Set GPU in both PCIE slots. Same result

Without GPU ( igfx )
- Ran OCCT with multiple profiles
- CPU runs at 100deg
- Power consumption 150W
- No issues. Does not turn off

- When I re-seated CPU after applying new paste, it won't post. Had to move RAM to different slots to make it work.
- RAM runs ok at 2133 as well as XMP @ 3000

- No issue with earthing
- Motherboard capacitors look ok
Post automatically merged:

Me and few members here had issues with Farcry and other games at start of this year and windows updates were the culprit. I uninstalled all crappy updates and since then havent updated my win 10 and games running butter smooth.
Was playing Far cry till yesterday. It's all ok, did not face any issue.
 
I think best way is to have it tested separately. Cause it might be possible that either rail on PSU feeding GPU has issues or GPU itself has issues.
 
I think best way is to have it tested separately. Cause it might be possible that either rail on PSU feeding GPU has issues or GPU itself has issues.
Agree

I'm looking for arrangements. Unfortunately Covid + Delhi rains are making things difficult.
 
I installed the GPU on my other desktop .. ( It was hell as the case is super cramped and need to be changed ) .

The GPU ran fine on boot. Fans working. Installed, stable drivers. And then ran Fur Mark at 400x300 ( Did not want to put too much load on Antec 450W psu ) ... Fur mark ran fine .. but my ups ( 600VA ) reported overload

Does this imply that it is the PSU that is having issue ?
 
I installed the GPU on my other desktop .. ( It was hell as the case is super cramped and need to be changed ) .

The GPU ran fine on boot. Fans working. Installed, stable drivers. And then ran Fur Mark at 400x300 ( Did not want to put too much load on Antec 450W psu ) ... Fur mark ran fine .. but my ups ( 600VA ) reported overload

Does this imply that it is the PSU that is having issue ?
PSU GPU 6 pin/8pin might have issues. Either try changing the cable if its modular or swap the psu.
 
Ok. But could it happen that without any external interference .. like pushing , stretching the cable/GPU pin might get damaged ?
Quite possible but also can happen due to psu age. Else even can happen out of nowhere, its electronics stuff eod.
 
Back
Top