M5A99FX Pro R2.0 random GPU freezes and system reboots

[dsterndorf]

I have a 4 year old M5A99FX Pro R2.0 with FX-8350, 2x4gb G-Skill DDR3 sticks, 1 TB WD Blue HDD, an ASUS Strix 960 GTX 2gb GPU, a 650 watt PSU in an Antec 900 case. This machine was upgraded to Windows 10 home 64bit from Windows 7 64bit three years ago. Within the last three months my machine began doing random reboots (with various errors displayed) and freezes. It would happen during gaming, work projects, and just idling. The problem would happen once every 10 days or so, but now happens every day and sometimes immediately after fully booting up. Sometimes the VGA led would light up red, other times leds were normal when the problem occurs

BIOS is up to date.

CPU temps are ok and run between 28C at idle to 48C under heavy load - measured both with CAM and HWiNFO.

GPU temps on the ASUS 960 GTX are also ok and run between 32C and 51C under load before crashes occur.

I saw errors in the Windows event log that seemed to indicate an HDD problem and ran chkdsk and sfc /scannow with no trouble found. The errors are things like Kernel Power, clock watch dog timeout or system thread exception not handled which lead me to believe my HDD was going bad or there were some flaky driver issues.

I did a fresh Windows 10 install on a brand new SSD and second new HDD that I had on hand and problem persisted.

I have replaced the existing 650 watt PSU as it was 6 years old, with a brand new 650 watt PSU, with no change just because I wanted to be sure power wasn't the problem.

I moved and removed memory with no change.

Memtest passes with both original sticks installed in their original slots.

I do not see any driver problems and drivers are up to date. For grins, I uninstalled and reinstalled with same results. I even rolled the driver back to the previous one also with no change in symptoms.

I pulled the ASUS 960 and borrowed an old EVGA GTS 450 and all runs fine. I threw in my even older EVGA 9800GTX and it also runs fine all day long. I do not have another rig to try the 960 in to test.

Just for grins I pulled off the CPU cooler, cleaned off the old paste and reapplied new paste. The machine measures 1 or 2 degrees cooler now, but I don't think I'm fighting a heat issue.

I want to believe that my GPU is bad as that would be the easiest solution, but my gut is undecided and I cannot rule out the mobo.

Is there anything else I can try or may have missed in my troubleshooting, or has my 960 simply bit the dust?

Thanks,

D

 

[DSzymborski]

You've done most of the things you should do (though I'd like to know specifics on the PSU just to be safe). That your 9800 GTX works is a big red flag here; it's more power-hungry than your GTX 960 is, so that it works fine suggest there's no power delivery issue in play here. I'm feeling very strongly that a failing GPU is the culprit here. Just out of curiosity, I'd also try running just one stick of RAM. Probably won't find an issue, but it costs $0 to just double-check this.

Before just going out and buying a new GPU, I'd probably do a full wipe and reinstall of Windows 10. If you haven't done that since upgrading, you likely have a really old install plus the in-place upgrade. I don't think it's the likeliest culprit given your PC's behavior, but it's good to exhaust all the possibilities that don't cost you anything.