Question Is my GPU damaged? - Crashing in Games, Benchmarks and making noise

kev.med

Prominent
Feb 5, 2023
14
0
510
Hey Guys,

so I had some problems with my PC in the last weeks and I am now wondering if my GPU might have taken a hit somewhere...

Games keep crashing and i get weird visual glitches when it´s under heavy load.

Below i made a video of said glitches when running a benchmark from 3d mark.
Visual Glitches when benchmarking
The red dots are what make think the GPU is damaged...

I´ve got a GTX 1070 Gaming X from MSI.

Has anyone seen anything like the visual glitches before? Could the PSU be the reason as well?
 

LanceF

Distinguished
Mar 6, 2012
3
0
18,510
I've encountered artifacts like that when overclocking the gpu. While it could be a PSU issue, I'd start with the easy stuff: overheating. Have you checked for dust (clean w canned air) and are your GPU fan(s) spinning? HWiNFO provides gpu temps and fan speeds.
 

kev.med

Prominent
Feb 5, 2023
14
0
510
I've encountered artifacts like that when overclocking the gpu. While it could be a PSU issue, I'd start with the easy stuff: overheating. Have you checked for dust (clean w canned air) and are your GPU fan(s) spinning? HWiNFO provides gpu temps and fan speeds.
Hey LanceF,

Thank you for your reply.

I had it out of the case yesterday and there was no dust visible at all. Im running furmark just now and the temps are at around 70° C.
Fans are spinning just fine, @ 2500 rpm at these temps.
With furmark there are no crashes at all .. i don´t know if this could mean anything...

If I try to test with 3dmark timespy the temps go up to 50 ° and then it crashes, which means it´s not a temperature issue i would guess?
 

kev.med

Prominent
Feb 5, 2023
14
0
510
If someone want´s the system specs. It has been changed recently so here are the specs before and after:

System Specs before:
- 500 Watt be quiet! Pure Power 10 CM Modular 80+ Silver
- MSI B250M MORTAR Intel B250 So.1151 Dual Channel DDR mATX Retail
- 16GB (2x 8GB) Corsair Vengeance LPX schwarz DDR4-2133 DIMM CL13-15-15-28
- Intel Core i5 7400 4x 3.00GHz So.1151 BOX
- be quiet! Pure Rock Tower Kühler
- 8GB MSI GeForce GTX 1070 GAMING X 8G Aktiv PCIe 3.0 x16
- 128GB Intel 600p M.2 2280 PCIe 3.0 x4 32Gb/s 3D-NAND TLC Toggle
- 1TB Seagate BarraCuda ST1000DM010 7.200U/min 6
- 2 TB Samsung SSD 870 QVO
- Fractal Design Define Mini C

System Specs now:
- 500 Watt be quiet! Pure Power 10 CM Modular 80+ Silver
- ASUS ROG Strix X570-E Gaming -- changed
- 16GB (2x 8GB) G Skill f4-3000c16d-16gisb -- changed
- AMD Ryzen 7 2700x -- changed

- be quiet! Pure Rock Tower Kühler
- 8GB MSI GeForce GTX 1070 GAMING X 8G Aktiv PCIe 3.0 x16
- 128GB Intel 600p M.2 2280 PCIe 3.0 x4 32Gb/s 3D-NAND TLC Toggle
- 1TB Seagate BarraCuda ST1000DM010 7.200U/min 6
- Hyte Y40 Midi Tower, Tempered Glass -- changed

bought all of the "old" components in April 2018.
 

LanceF

Distinguished
Mar 6, 2012
3
0
18,510
I appreciate you upgraded the hardware for a reason; I assume your newer CPU is yielding higher frame rates than the older hardware combo. Furmark and Time Spy test the GPU differently. You could try undervolting your GPU/ lowering clocks with MSI Afterburner and see if you can achieve stability with Time Spy. You could use NVIDIA Control Panel and set a max frame rate, too (or do the same in your game) to see if that yields better in-game performance/ stability (overall). (If the PSU worked before, I assume it is good enough now: No reason to believe otherwise.)
 
Last edited:

kev.med

Prominent
Feb 5, 2023
14
0
510
I appreciate you upgraded the hardware for a reason; I assume your newer CPU is yielding higher frame rates than the older hardware combo. Furmark and Time Spy test the GPU differently. You could try undervolting your GPU/ lowering clocks with MSI Afterburner and see if you can achieve stability with Time Spy. You could use NVIDIA Control Panel and set a max frame rate, too (or do the same in your game) to see if that yields better in-game performance/ stability (overall). (If the PSU worked before, I assume it is good enough now: No reason to believe otherwise.)
I did set a FPS Limit, that unfortunately didnt change anything. The PSU was most likely not at fault (I have gotten another one to check with the same results).

I don´t really get how I can try to solve my problem by undervolitng tho.
As far as I understand that topic undervolting may result in a more unstable solution. I have watched a video about it and the guy in it says that you are supposed to lower the voltage until games become unstable before slightliy raising it again. (The video i watched:
View: https://www.youtube.com/watch?v=KPR06CxysMw&
)
How would you recommend I undervolt it to make it more stable?

Thank you very much in advance for your answer :)
 

LanceF

Distinguished
Mar 6, 2012
3
0
18,510
Sorry I haven't been back sooner. The idea behind my suggestion to limit FPS and/or undervolt is that, while the board doesn't seem to be overheating, perhaps a component on the PCB is overheating which is causing the crashes. Reducing FPS and reducing voltage are two ideas to reduce power requirement; if a component (whose temp isn't being measured) is failing, one of those strategies may keep that part cool enough to keep it running. As a 3rd prong, I guess you could also max-out fan speed (again using MSI Afterburner).

From your new-to-you parts list (I assume they were all purchased used), while I don't see a red flag, you could reinstall your old mobo/parts and see if the problem is only with the new hardware. Of course, everything I'm suggesting assumes your OS and drivers are up to date (and you're not OCing your CPU, RAM or GPU or, if you are, you have tested those somehow... and that you "loaded optimized defaults" (and perhaps RAM XMP profile) in BIOS). Sorry. Long list! Cheers!
 

kev.med

Prominent
Feb 5, 2023
14
0
510
Sorry I haven't been back sooner. The idea behind my suggestion to limit FPS and/or undervolt is that, while the board doesn't seem to be overheating, perhaps a component on the PCB is overheating which is causing the crashes. Reducing FPS and reducing voltage are two ideas to reduce power requirement; if a component (whose temp isn't being measured) is failing, one of those strategies may keep that part cool enough to keep it running. As a 3rd prong, I guess you could also max-out fan speed (again using MSI Afterburner).

From your new-to-you parts list (I assume they were all purchased used), while I don't see a red flag, you could reinstall your old mobo/parts and see if the problem is only with the new hardware. Of course, everything I'm suggesting assumes your OS and drivers are up to date (and you're not OCing your CPU, RAM or GPU or, if you are, you have tested those somehow... and that you "loaded optimized defaults" (and perhaps RAM XMP profile) in BIOS). Sorry. Long list! Cheers!
Thank you very much for your answer! I also didn´t have the time to come back to this topic :)

I tried putting the fans up to 100 % at all times, reduced the power limit and set core and memory clock at -100 MHz.

Aaaaaand the benchmark finished!

I then step by step undid the changes and what i ended up with is fans with a custom fancurve as before, the power limit at 100 % and the memory and core Clock each at - 70 MHz.

I will now try to stress test a bit more but for now thank you very very much for your help. I was trying way too long to figure out a solution!!