What's happening to my GPU

Peddeh

Member
Jun 19, 2013
33
0
16
It's a long story but I'll try to keep it short and hopefully you'll all understand my gibberish. This has to do with my video card, the R9 290X Lightning from MSI.

This all started about 4 weeks ago, when I picked up BF1 and noticed that this was happening https://imgur.com/QplUErV - followed by black screening. I noticed that underclocking by ~20% solved the issue, although just in BF1, as I later found out in Titanfall 2 it would black screen regardless.

Fast forward: I filed a direct RMA with MSI. Sent it off and within a day or two MSI shipped a replacement card back to me, a refurb, which I've had in my possession since the 13th.

Card has been great up until today, no problems what so ever. Only until today, after I experienced a BSOD during a BF1 MP match, which I presumed was to do with my overclock on the CPU as I had upped my multiplier a notch and vcore and tested it under p95 (which seemed to be fine for an hour), so I just knocked it back down. Booted right back into windows, jumped back into the game, 5minutes later and the artifacting kicks in, as I saw on my previous card.


Is this a coincidence, could my PSU be involved?

I'm absolutely stumped :(

My system specs:
4930K @ 4.4GHz (1.310vcore) - Stable for 2 years now (it was at 4.5 when I experienced the BSOD in BF1)
Rampage IV Black Edition
16GB Corsair Dominator 1866MHz
MSI R9 290X Lightning (Stock)
SuperFlower Leadex 1000W Gold
 

EXCellR8

Diamond Member
Sep 1, 2010
3,979
839
136
First step is to abandon all OC and revert to stable defaults, simply to rule out any underlying issues with hardware. Remove all but 1 memory DIMM also. SFC is a power supply OEM, so that's good, but it is still possible that the PSU is involved in the crashing.

Once your board is set to defaults, check to see if there are any BIOS updates available. Even simple stability updates can affect gaming, so it's at least worth a look.

Make sure Origin and your games are up to date, along with the drivers for the 290x and possibly check for signs of overheating. Cards that get hot (290X is certainly no exception) can fail once they reach a certain temp--or show visible artifacting or trigger BSOD.
 

Denithor

Diamond Member
Apr 11, 2004
6,300
23
81
As said, turn everything back to stock first. Next up, run Furmark to torture test your GPU. If you get artifacts/bsod/etc you probably have a GPU problem.

PSU problem is much harder to isolate, unless you have a backup you can swap for and retest, error free with new PSU indicates problems with the current one.
 

casiofx

Senior member
Mar 24, 2015
369
36
61
Just to share my experience of black screening of my R9 290 tri-x last time.

I had been running my i5 4440 and R9 290 with Cooler Master Extreme 2 625. Not a good power supply but it powered my computer fine for 18 months no problem, even when I overclocked the R9 290 to +200mV and 1230Mhz. Then it started to black screen on stock settings and even when the GPU is idling on windows desktop. It would totally black screen randomly around 5-30 mins of turning on the computer, once it black screened you have to turn on the power plug to turn it on again. Thought was the GPU's problem and I nearly sent it for RMA.

Just when I saw some cheap Seasonic X-750 on the computer store along with a SIlverstone mid-tower, I bought it to replace my matx casing. Then, no more black screen, the system is totally stable... It was never the problem of the GPU, but the PSU that had deteriorated components. The PSU now happily powers my sister's computer with i5 4440 and 1050Ti (previously AMD Athlon II X4 635 and HD5830, no stability problems)
 

Peddeh

Member
Jun 19, 2013
33
0
16
Just to give an update.

Rolling back to stock settings on everything didn't help, that was one of the first things I tried.

Unfortunately though the card now appears completely dead. I ran OCCT's PSU test on it (which puts everything under load). Within ~15s, the system black screened and the GPU fans ramped up to 100%. After resetting the PC, it won't post. No matter what slots the card is in. I'm using an older HD 5450 at the moment with a new card on the way, a 1070.

Going to try it with the 1070 and hope I won't have to return both the card and the PSU (still under warranty).
 

ConsoleLover

Member
Aug 28, 2016
137
43
56
Seems like a bad graphic card. Also you seem to be a newbie on OC, so for all final overclocks you need to run a stress test for 24 hours. That is the only way to know for sure if the clocks are stable.

I also wouldn't put the 1070 in on a possibly faulty PSU. Best way to check is to try some reserve PSU, or borrow a psu from a friend or whatever and test it.
 

Peddeh

Member
Jun 19, 2013
33
0
16
Update: Okay, so I took a bit of a risk. My 1070 arrived, and being impatient, not wanting to wait for a reserve PSU decided to test it on the current PSU. Luckily it didn't go bang in OCCT's PSU test. Seems to be stable (only left it for about 10m, which is alot longer than the ~15s the Lightning lasted). No artifacting, no abnormal temperatures. So far so good!
 

Bacon1

Diamond Member
Feb 14, 2016
3,430
1,018
91
Well if its not the PSU, make sure you send the card back to MSI for another RMA and say the one they sent was faulty.
 

bystander36

Diamond Member
Apr 1, 2013
5,154
132
106
I've gone through a similar experience with a bad PSU. After one GPU went bad, and RMA'ed it, the new GPU started to go bad a few months later. This time I replaced the PSU, as when I stopped using SLI, the problems went away, and both cards worked fine.

The PSU might power the new GPU for a while, but if it is the faulty piece, it will cause your GPU to start failing in the future.