Testing a video card *updated*

[Fanatical Meat]

My PC is having trouble starting up, it does to a black screen it started about 6 weeks ago and now its pretty regularly happening. I just had to restart my PC 8 times to get it working. I suspect its the video card, I do not have an on board option to use. Does a test exist to see if its the card, like a benchmark kind of thing?

I have an Asus R9 280X

Secondly if I get a new card would going to a GTX950 be a silly idea? The machine is getting pretty old and in my computing experience I've owned 5 AMD/ATI cards 2 of which have failed around the 20 month mark (assuming this one is dying). I find that failure rate pretty irritating. I don't want this to be a bash AMD thread.

I just did a clean Window 10 reinstall two days ago.
Q9650 w/8GB & ssd.
1680x1050 monitor

 

[Smoblikat]

When the computer boots with a black screen, do the lights above caps lock/numlock light up? Can you toggle them? Do you hear the windows login sounds eventually?

There are a few things to try first:
Reseat power connectors to GPU
Reseat the GPU itself
Reseat memory
Use a different card in the computer and see if its doing the same thing
Take your card and put it in another computer and see if it does it there too

If you can get it to boot into windows, run GPU-z and monitor temps, then run furmark/OCCT/Kombuster/Unigine heaven/whatever other stress testing program you want. Pay very close attention for artifacts, which could appear as random dots of color or black lines/flickering.

The most important thing to do is put the card in another computer and see what it does. If it happens in another machine, try baking it, ive done that many many times and it has helped.
http://lifehacker.com/5823227/save-dying-video-cards-with-a-quick-bake-in-the-oven

I usually do 375F for 11 minutes.

EDIT - to answer your second question, a GTX 950 seems like a decent replacement for a 280:
http://www.anandtech.com/bench/product/1591?vs=1670

That shows a 380 being slightly faster. Ive owned a lot of nvidia and ATI cards and never had any consistent problems with either brand.

 

[Fanatical Meat]

Thanks Smonlikat, you sound pretty knowledgeable in this topic. Do you agree it sounds like a video card problem?

 

[UsandThem]

Thanks Smonlikat, you sound pretty knowledgeable in this topic. Do you agree it sounds like a video card problem?

He wrote a great post on steps you need to take to determine if the problem is your video card or something else.

 

[Fanatical Meat]

He wrote a great post on steps you need to take to determine if the problem is your video card or something else.

Thanks for the input. I was looking for an opinion.

 

[UsandThem]

Thanks for the input. I was looking for an opinion.

If you are looking for an opinion without testing, then yes, it is your video card.

 

[Runequest2]

It could also be your Power supply.

So doing what was suggested will help figure out if it is the GPU.

 

[Smoblikat]

Thanks Smonlikat, you sound pretty knowledgeable in this topic. Do you agree it sounds like a video card problem?

It could be the video card, it could also be the PSU, or even the motherboard. I know on older ASUS boards with multiple PCIe X16 lanes it likes to change which one the video card should be in. I had a P5Q pro turbo and would occasionally have to move my card to the other slot to get it to post video. No idea why that happened, but thats just what ive experienced. Id try a few of the steps I listed above and go from there. Until you do some more testing, it could be any number of things.

 

[IEC]

It could be any number of things.

Test to eliminate one possibility:
Check to see if ASUS has a newer vBIOS for your card. Sometimes on the 200 series Radeon cards they need a little extra voltage (e.g. +20mV) after a few years to be 100% stable. I had it happen to me on a Sapphire card and they sent me a updated BIOS that gave it +25mV on the core and fixed black screening.

 

[AnonymouseUser]

That's an old setup, so I would be looking at the PSU or motherboard before the much newer GPU.

 

[Fanatical Meat]

Well good news the current owner of my old GTX260 is willing to try a swap to see what happens

 

[Smoblikat]

Well good news the current owner of my old GTX260 is willing to try a swap to see what happens

Cool, let us know how it goes.

 

[Fanatical Meat]

I forgot to add its only happened once when the computer was on. I alt-tab out of a game and had some graphics error, something about the high density display. Every other time its happened from start up or when the PC was starting after sleeping.

 

[tweakboy]

Play games,,,, real world my friend...

 

[Fanatical Meat]

Play games,,,, real world my friend...

Thanks for the assistance....I think maybe this is sarcasm?

 

[cusideabelincoln]

In my experience, the Crimson drivers under Windows 10 are not very good for the 280x. I had issues with them, and of course (even though I disabled automatic driver updates) for some reason Windows 10 update would install the Crimson drivers. I did a clean sweep of the drivers three stinking times. So I'm sure that, along with the older system you're running, is the potential problem.

 

[Stuka87]

To answer your question about the 950, that would be a huge downgrade from the 280X.

As old as your machine is, kind of sounds like you may have some dried up caps either on the mainboard or the PSU. It certainly could be the GPU, but without doing some actual testing, its impossible to know.

 

[Headfoot]

So when you say multiple cards have failed, your distrust is 100% entirely misplaced by aiming it at the brand of GPU which has almost nothing to do with failure rate. You should be looking at your power supply or the OEM brand used.

Simply put, AMD or nVidia has nothing to do with failure rate. Using a crappy/old/out of spec PSU that keeps frying your cards does.

 

[Fanatical Meat]

Corsair 620 power supply
Other failed card was in a old dell machine I had

 

[Stuka87]

Corsair 620 power supply
Other failed card was in a old dell machine I had

Still, the failures had nothing to with AMD. And everything to do with whatever brand the card was (ie: ASUS, MSI, Sapphire, etc)

How old is the PSU? And which model?

 

[Rifter]

Im leaning towards PSU or Mobo as well for what its worth. testing is only way to know for sure.

 

[Fanatical Meat]

Here's the update:

I checked the cards seating it was good however I didn't remove the card
I was planning on doing the swap with my old gtx260 tomorrow. When I took the problematic 280x out it rattled, I noticed there is a shroud around the fans and heat pipe setup that was very loose. I found two "clips" one was very loose and the other was bent. Looked like this held the heat pipe down to the chip. I tightened them and rattle disappeared. I reinstalled and I was able to restart 5 times without incident.
I wonder if the chip was overheating but I don't understand why it would be a problem when it powers on from being shut down. I don't have any thermal compound but I guess I should take it apart and reapply some.

Edit: The clips were mount type things for the screws

 

[mikeymikec]

It took a fair while to decide that my 750ti was faulty. The symptoms were:

Intermittent graphics corruption followed by TDR during gaming
Intermittent issue resuming from sleep mode (sometimes the computer would be stuck between S3 and anything showing on the screen, or BSOD)
Intermittent issue after resuming, graphics corruptions in video playback

After testing everything else, I swapped in an older graphics card, no problems. That was the clincher as far as I was concerned.

 

[AnotherGuy]

Its so great when people find the problem themselves...
See I wouldn't blame the brand of the vidcard so quick, because if they made cards that last only 20 months then they would be out of business long time ago.

 

[Fanatical Meat]

I still don't understand how a loose cowl and cooler would cause it not to start from being off overnight. I've restarted more times today and the problem hasn't occurred.
How could the chip get too hot within seconds?