OC Issue: Three tries to boot, then PC runs fine?

[hyrule4927]

I've been having a weird problem lately. When I boot my PC, it always has to try twice. Right after I hit the start button, it will start up, shut down before the post screen appears, then start again successfully. Then, at Windows login (Windows 8 preview), none of my USB devices will work, and I'll have to hit restart once before I've got a working keyboard to log in with.
It is somehow overclock related, since I don't have any of these issues at stock clocks, but I'm not sure what to do about it. Once I'm finally into Windows, everything runs great. I can run Prime95, StressCPU2, and any game for hours without a hitch, and I can run Folding@Home for days without any issues, so what's with these weird problems at startup?

System specs are in my sig. Running the CPU at 3.8 GHz right now because I like to keep my temps below 50C while folding 24/7.

 

[SickBeast]

Are you sure that your overclock is remaining after all those reboots? Sometimes if your system can't handle an overclock it will reboot itself a couple of times and revert to the default values.

 

[hyrule4927]

The correct core speed is showing up in POST, BIOS, and Core Temp even after all the reboots. I can't check with with CPU-Z because it just freezes in Windows 8 and HWInfo seems incapable of displaying proper speeds in Windows 8 as well (shows 258 MHz for my CPU and 54.4 MHz for my RAM right now, and equally weird values at stock settings).

Also, if I shut off my computer then start it up again quickly, there's usually no problem, but if I shut if off for 10 minutes or so then turn it on again, the reboots always happen.

I think I might try bumping my voltage a bit and trying again. It'll be a little while before I can report back, since I have to let it sit for a bit to make sure I have consistent results. Any other suggestions are welcome.

Edit: Bumped the voltage and had even more boot cycles, although I don't know if the two are related. Ugh . . .

Edit 2: Happened at stock settings too. Looks like somethings wrong with my system, but no idea how to diagnose it.

 

[SickBeast]

The most likely culprits are usually the motherboard or else the PSU. PSUs die over time. They have a shelf life basically.

You could always RMA your motherboard if you don't have spare parts to test with.

Try disconnecting everything from your system aside from your CPU, GPU, and 1 stick of ram and see if it posts ok.

 

[hyrule4927]

Okay, letting my system sit before I start again. Reset the BIOS as well for good measure. Knowing me I could have screwed it all up with some little setting change or another.

And I really hope it isn't the PSU. Bought it about two weeks ago because I'd heard so many good things about Seasonic . . .

Sucks if it's the MOBO too though. Didn't want to get a new one until I could afford a 2600K.

Edit: One successful boot. Maybe my CPU is getting to old for the OC's I've been running, unless some other BIOS setting was out of order? Going to let it sit for 10 again and see if I can get two successful boots.

Edit 2: And . . . another failure.

 

[bononos]

I had a failing mb which did alot of strange things. Multiple tries needed before booting sometimes if booted cold in the morning. And if it didn't, it almost always only took a 2nd jolt before it responded favourably. During the times it failed to boot fully it died a second after I punched in the power button before the bios got a chance to chatter. And when it booted it worked ok.

I changed out the psu because the voltages were off a little (still within spec) but it didn't help the boot issue one bit so I guess the issue was with the mb. It was quite a while back and it used electrolytic caps so I guess they had degraded to the point of needing a little prodding to wake up.

 

[hyrule4927]

Hm, I'd rather it be the mobo than the power supply to be honest. Fooled around with some wires, had another successful boot. Guess it's just going to be a mixed bag. As long as it actually runs in the end, I can live with this for now. It's my hobby computer, not my primary PC. Just hope nothing dies outright before I can afford the upgrades I want . . .
Was intending to use some rebate cash I've got coming to pick up a second 6850, but I'm going to have to seriously consider getting a 2500K and motherboard combo. I'd really rather wait until I can afford a 2600K or even hold out for a Haswell i7 (if they keep that name), but an involuntary upgrade might be on the horizon.

Can't think of anything else to test for now. Tomorrow I'll try hooking up my old CX430 to see if my PSU is the problem. Worst part is how long it takes to test since the problem only manifests itself on cold boots.

 

[bononos]

I doubt that your psu is faulty since Seasonics are very good.
The fact that you're overclocking and running it 24/7 means your pc is going to wear out sooner.
How long have your been running your pc like that?

 

[hyrule4927]

I've only been folding for about 2 months, and the PC's had some intermittent breaks in there (haven't run it nonstop for more than a couple days at a time as far as I can remember). Bought the CPU and motherboard used off of the NBR forums marketplace 3 months ago, so I can't speak for the kind of use they saw before then. Was kind of a "just for the heck of it" budget build, but buying used was still probably a mistake.

And the 4.4 GHz in my sig is really just the highest I've gotten it stable at and run some benchmarks. Actually been running it at 3.8 GHz the majority time with the core voltage around 1.32 (Intel's recommended max is 1.4). Temps are rarely above 50C.

Edit: Tested with the CX430 and the issues persist, but I think I'm on to something. If I'm right I'm going to feel like an idiot, but this would also be the best case scenario . . .

I realized that my graphics adapter setting in BIOS was defaulting to PCI instead of PCI Express. Maybe this little error confused things enough to cause all my problems? Set it to Onboard and the system booted fine with video output from the HDMI on my mobo I/O panel. Going to test again to see if I get a failed boot with it set to PCI. Then I'll set it to PCI Express, and if that works, I think I'll have solved my problem!

Edit 2: Nope, back to square one. Going to get some sleep, then more testing tomorrow. Going to see if I can manage multiple successful boots with my BIOS set to Onboard graphics, which would suggest an issue with my GPU or PCI Express slot?

Maybe I can borrow my friend's 550ti for a quick test with a different GPU. The answer will probably be "no" though . . .

 

[bononos]

Its a good idea to see if integrated gfx instead of the 6850 fixes the booting issue. I think the mb can't handle the juice that the 6850 needs that well on a cold boot.

 

[hyrule4927]

Now it's not happening at all. Which is a good thing, but it also means that any correlation I thought I identified might have been a fluke. Any more ideas?

Edit: Let it sit for 45 minutes instead of 10 and had another weird boot with PCI Express. Waiting another 45 before another boot with Onboard.

Edit 2: Looks like the GPU issue is still most likely. And should motherboard power input really be a limiting factor? The GPU has it's own power cord after all.

 

[bononos]

It still takes some juice of the pcie slot. 16x pcie slots are supposed to deliver 75watts max (higher for newer pci-e specs).

 

[hyrule4927]

Gotcha. Now does this actually indicate a defect or impending doom of the motherboard, or is it simply an inherent limitation of an older motherboard?

 

[bononos]

I can't think of any reason for it to be an inherent limitation. I would just make regular backups and ride it out until upgrade time.