Question Red lines, blue screens, crashing, screen going black

[Belch45]

Specs:
1660 6gb
16gb ram (8x2)
Ryzen 7 1700x
Msi x470 Motherboard

It started with my screen going black and me not being able to hear anything. When I shut off the pc and restart I get those red lines and most of the time the resolution would turn down significantly. If i had to guess I would say like 480p or less. Sometimes when the screen goes black I can still hear and sometimes its a different color. One time it was gray. When the pc is working I can sometimes see brief red lines somewhere on the screen and have seen red squares on the screen as well. I tried redownloading the NVIDIA drivers and even tried switching to studio drivers (I used the clean install option) In the link below you can see the error messages Ive gotten from the blue screens and the red lines on startup.

The two different blue screens+ red lines

 

[rgd1101]

brand and model of the psu?

check the gpu temp?

 

[Belch45]

Corsair RM 850x, I havent checked the temp ill try that right now.

 

[Zerk2012]

Monitor, cable, video card.

 

[Belch45]

Graphics card: NVIDIA Geforce GTX 1660 6gb, HDMI, Acer Gaming Monitor 23.6 curved ED242QR Abidpx 1920x1080 144 hz

 

[Belch45]

I was looking at the gpu temperature and it was running around 30c. I went to grab something when I got back the screen went black except this time I can see my mouse cursor and the search bar when I hit the windows key.

 

[DimkaTsv]

I was looking at the gpu temperature and it was running around 30c. I went to grab something when I got back the screen went black except this time I can see my mouse cursor and the search bar when I hit the windows key.

Idk what i am thinking, but kinda looks like memory corruption...
Question is, is it GPU memory going wild or i am wrong... Maybe it is RAM, but RAM corruption leads to much bigger problems usually
Do you have your 1660 OC?

Also can be driver errors, some incompatability, etc...

 

[Belch45]

I think at one point I put OC on my graphics card via MSI Afterburner when I was mining ethereum. Im not sure if its still overclocked or not but when I was trying to fix it I reset the settings on MSI Afterburner.

 

[DimkaTsv]

I think at one point I put OC on my graphics card via MSI Afterburner when I was mining ethereum. Im not sure if its still overclocked or not but when I was trying to fix it I reset the settings on MSI Afterburner.

Nah, it it was just OC problem you would've saw it almost instantly after setting it up... Especially with visual glitches that bad
Usually GPU RAM corruption from OC can be noticed as some pixels (not most of screen) are going blank or white for really short time. It is not lethal symptom ofc, just signal "Hey you should lower that memory frequency a bit, otherwise...

 

[Belch45]

Yeah it certainly wasnt the ram sticks. Do you know if I could use a gddr2 graphics card with my motherboard? (MSI x470 gaming plus)

 

[Karadjgne]

Nah, it it was just OC problem you would've saw it almost instantly after setting it up... Especially with visual glitches that bad

Not true. With voltages/power limits set too far over the edge of safety, but still within stable tolerances, the level of electromigration and transistor burnout is magnified logarithmically. You'll be good for a minute, but over time that irreversible damage is spread until it starts to affect operation. There's 6.6Billion transistors on a 1660 I believe, can afford to loose more than a few before the affects are noticed.

Whether it's VRM or VRAM burnout is a moot point, the card is toast.

 

[DimkaTsv]

Not true. With voltages/power limits set too far over the edge of safety, but still within stable tolerances, the level of electromigration and transistor burnout is magnified logarithmically

I guess you're right... But is there actually available for public way to set limit over safety edge long-term without noticing?
Also, i am not talking about GPU overclock rn, just increasing memory frequency.
Of course if you set it up to unbelievable amount you can instantly become a loser in all meanings there. But are there any person that will do burst overclock to unreasonable amount and not move slowly but surely looking on instabilyties
Yes there is also degradation that must be taking in account... But i doubt that non-professional overclocker will actually increase voltages over recommended or power limits that he cannot cool down...
On other side... There are all kinds of people existing, so there may be such that will do that stuff

 

[Karadjgne]

Yes. Msi Afterburner, Evga Precision X, any of those OC software titles can and do have sliders to raise power limits, raise voltages, push ram. Most cards nowadays are already close to power limits, adding another 15% or 20mv can easily put them beyond the physical limits of the silicon, regardless whether the sliders are maxed or not. Silicon lottery.

It's a mining used gpu. That's a game changer. Power outputs are vastly different there as are stability outputs since literally half the card goes basically unused.

When OC a gpu, you'd take just clocks as high as could go stable, reset, take vram as high stable then reset, then both a little lower. You'll not get both pushed to the stable limits simultaneously as that puts voltage requirements far above healthy. Balancing act. Being a mining gpu, you'd push vram limits, but clock speeds are secondary, but even so, that's a high degree of use and voltage demand.

Which if done wrong can drastically shorten usable lifespan. Most professional mining systems use undervolted cards, not overclocked/overvolted cards as there's no benefit to shortening usable lifespan with OC.