• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

GPU - Black screen crash

Associate
Joined
12 Oct 2011
Posts
23
Hi, I'm looking for any advice on how to test a GPU for faults (Sapphire R9 390 8GB) about 2½ years old, so out of warranty. With graphics card prices being a bit mental for anything capable of running 1440p well, I'd like to try and see if I can find a work around for a month or two so I can get some funds together to replace it and the rest of the PC.

Basically, while gaming, the machine frequently fails to a black screen at times - usually while something is loading (typically a 'mission end' screen or 'map/area' transition).

I have swapped the card with that from another machine (with identical card) and the fault followed it.

I've put the Memory through a single Memtest pass - for each stick, both pairs and all four - they passed.
I've also re-installed windows (on new SSD) and running on low graphics seems to have stabilised things somewhat. The crashes are now just occaisional rather than regular occurances.

Any ideas greatfully received.

For info, most of the rest of system is 6½ years old, so possibly getting a bit long in the tooth - i5-2500K (now back to stock), ASUS P8Z68-V, 16 GB RAM (4x 4GB), and 750W Corsair TX PSU.
 
Hi, I'm looking for any advice on how to test a GPU for faults (Sapphire R9 390 8GB) about 2½ years old, so out of warranty. With graphics card prices being a bit mental for anything capable of running 1440p well, I'd like to try and see if I can find a work around for a month or two so I can get some funds together to replace it and the rest of the PC.

Basically, while gaming, the machine frequently fails to a black screen at times - usually while something is loading (typically a 'mission end' screen or 'map/area' transition).

I have swapped the card with that from another machine (with identical card) and the fault followed it.

I've put the Memory through a single Memtest pass - for each stick, both pairs and all four - they passed.
I've also re-installed windows (on new SSD) and running on low graphics seems to have stabilised things somewhat. The crashes are now just occaisional rather than regular occurances.

Any ideas greatfully received.

For info, most of the rest of system is 6½ years old, so possibly getting a bit long in the tooth - i5-2500K (now back to stock), ASUS P8Z68-V, 16 GB RAM (4x 4GB), and 750W Corsair TX PSU.

Seen as the fault followed the card you could try stripping it down, taking off the cooler, giving it a good clean, fresh paste and seeing how it gets on. It might be worth pulling the clocks back slightly as well and seeing if that stabilises it.
 
When the "machine frequently fails to a black screen", how does it "fail"? Is it a complete lockup? Does sound continue in the background? Looping sound?

Wonder if the 390 has the "black screen bug" that 290s had... I thought that they were all resolved by the 390 series but if not a GPU BIOS update generally fixed that. Doesn't make much sense as to why a card would develop that bug though.
 
Thanks for the replies. Gives me a few new directions to work from.

When the "machine frequently fails to a black screen", how does it "fail"? Is it a complete lockup? Does sound continue in the background? Looping sound?

Screen just goes black, sound stops, USBs still pass power but no keyboard response (needs powered off on the front panel).

Wonder if the 390 has the "black screen bug" that 290s had... I thought that they were all resolved by the 390 series but if not a GPU BIOS update generally fixed that. Doesn't make much sense as to why a card would develop that bug though.

Try under clocking it to slightly below reference speeds, i had a 280x that was only stable at reference speeds. Also see if there is a bios update for it. Also monitor the temps for it.

It has only recently started doing it. I'd been on the version of the crimson Adrenaline driver (not the optional) for months - only updated to the most recent with the re-install. Will take a look to see if there are any BIOS downloads about. Having difficulty monitoring temperatures (can HW monitor create a log that will survive a reset?) but the fans don't noticably spin up before/during the black screen.

Interesting about dropping the clocks, I hadn't thought of trying that - should reduce the stress on the card much like running graphics in low settings does?

Seen as the fault followed the card you could try stripping it down, taking off the cooler, giving it a good clean, fresh paste and seeing how it gets on. It might be worth pulling the clocks back slightly as well and seeing if that stabilises it.

Kind of what I was expecting to have to try, but wanted to see if there were a few easier (even if they are a lot less fun) options than taking it apart first.
 
Thanks for that link Rikk82, makes for interesting reading alongside the other black screen crash threads or familair sounding tales of woe that have cropped up in the forum. What did you use to lower the memory clock? I was playing around last night trying to monitor things and reduce the momory frequency with Wattman, and the rate of crashing went up to amazing new levels (should have done something useful like go to the pub instead).

Wattman has just two states for the card - 150MHz and 1,500MHz - and it only allows me to increase the high frequency, but will not let me reduce it. I'm just going to try remove Wattman and see if Afterburner will work instead.

I have found some old Arctic Silver paste in my bits box and it is still squishy, I'll follow Vince's advice and break out a screwdriver if I don't get anywhere with killing Wattman.
 
Back
Top Bottom