PC Crash overnight; BIOS corrupted on Gigabyte 990FXA-UD3

[neiby]

I built a new PC a couple of weeks ago with a Gigabyte 990FXA-UD3 mobo and an AMD FX 8350 CPU. I have two nVidia GTX 460s, an SSD, a Blu-Ray drive and a couple of HDDs. The power supply is a Thermaltake SMART 850w. The PC has been running great for a couple of weeks. No problems. It was running fine when I went to bed last night. I had left it on, but it was off when I woke up and would not turn on. It was completely unresponsive.

I removed the power cable for about 30 seconds then plugged it back in, after which the PC started up. The interesting thing is that the main BIOS was corrupted and it had to recover from the backup BIOS. What would cause that sort of behavior? Some sort of transient power problem or something? It's very odd. I don't think I've ever seen this before.

 

[sin0822]

I would update your UEFI to the latest version. Sometimes the BIOS can get corrupt and the dual BIOS feature will kick in at the first sign of corruption, it might not even be needed but the corruption threshold might be a bit low on the BIOS version. Think of it of like how when you pull out a USB drive from windows many times without safely disconnecting you get that "please scan for errors" message because Windows thinks the drive might be corrupted or how sometimes windows might want to run chkdsk if you shutdown windows incorrectly over and over. Dual BIOS works kind of like that. Overclocking will also cause BIOS corruption, but a bad PSU might as well if there was a power surge and it messed up the CMOS(which is okay). That is what dual BIOS is for and IMO its a pretty good feature for overclockers.

 

[neiby]

I updated my BIOS right after building the PC, so I guess I'll chalk this up to some weird power surge that happened over night. The PC is connected to a surge suppressor, but it's kind of old. Who knows if it even really works anymore. Maybe something got through and caused the mobo to wig out a little. I'm sure happy to have a dual BIOS mobo now. That's pretty slick.

 

[neiby]

I went ahead and updated UEFI again to make sure I had the latest version. Things are looking good for now. Hopefully that was just a one-off glitch.

 

[VirtualLarry]

Is that board + CPU overclocked? That could cause some weirdness.

 

[neiby]

Nope, not overclocked but I was considering it.

 

[Kougar]

Even without overclocking have you run stability checks anyway to be sure everything is rock stable? I've seen plenty of rigs where a RAM setting is off and it can cause stability issues.

 

[neiby]

Even without overclocking have you run stability checks anyway to be sure everything is rock stable? I've seen plenty of rigs where a RAM setting is off and it can cause stability issues.

I've run 3D Mark and PC Mark 7, but those are obviously more benchmarking than stress testing. I've run prime95 to check how well my cooling is working. My CPU gets up to around 48C, maximum, and usually it's not even that hot. When not under load, it drops back down to low to mid 30s.

I suppose I could run prime95 for a longer period and then maybe run MemTest. What else would you recommend?

 

[Kougar]

Those are the main two, although OCCT also offers Linpack load testing and AVX Linpack testing for those that want to cover all the angles.

Without overclocking a single pass of Memtest should be sufficient. I'm fuzzy on how long Prime is recommended for these days, but since you've already run it for 5-15 minutes to check temps I wouldn't even worry about it.