Peculiar motherboard malfunction...

[Grall]

My current PC rig has always been a bit wonky, pretty much from day one, where I have had to reboot it after about half an hour or so every time it's cold-booted, or else it starts to stutter and soon bluescreens a while after starting any 3D application. After warming up, it works as expected, but it needs that first restart for whatever strange reason or no amount of warming-up will help. Weird...

As of late, this behavior has extended to rather slow POST during boot-up, literally taking 5-10 times longer than normal. Even navigating the BIOS setup is infuriatingly slow, with several seconds inbetween screen updates for each keypress.

Today, my PC decided to take this one step further. Instead of POST taking maybe a minute to complete it now took somewhere between 5 and 10 minutes (!), and after booting into windows any screen operation involving the GPU ran in slow-motion.

Weird thing though, if I powered off the PC with the power switch during POST and then booted it up, it behaved normally during POST (which is, slow as hell compared to most PCs, but not ludicrously-1-minute-plus-slow) and also in BIOS setup, but after leaving BIOS setup and restarting it was back at ultra-superslow speed.

However, after another restart the PC had warmed up sufficiently, and now runs normally again. Crazy!

I'm aware this most likely means my mobo is on the verge of failing, but I'm not keen on turning it in for warranty replacement this close to TES6:Skyrim's launch, because it'll most likely take a couple weeks to get a replacement board... Oh, if only Apple hadn't been giving gaming the evil eye for 30 years before seeing the light, maybe most games would have released both for windows and for macos...

 

[zed]

Simple thing you can try
Try booting it up in linux.
Ive seen it give much more information with hardware errors in the past, windows just ignored them and wouldnt start, eg it might be the memory. If so then linux when it starts should say 'hey youve got faulty memory' windows & most like mac os dont want to burden the user with this knowledge, so just wont start up

 

[digitalwanderer]

I hate wonky mobos like that, I've had a few. You have my sympathies.

What make/model is it and have you added any USB things lately? Also, what viddy card?

 

[Grall]

Mobo is Asus Rampage 2 Gene, with Sapphire Radeon 6970 viddy boards (reference design).

No new components added in ages. I did flash the BIOS the other week, but it made zilch of a difference in the way the PC behaves.

 

[almighty]

Remove all USB connections, Use 1 stick of RAM.

- Use the first revision BIOS and work up from there.
- Remove and re-seat the CPU ( Or even try a different one )
- Try different RAM and use different DIMM slots
- Run memtest86 in DOS to check for RAM errors

If you still can't find the fault then send in for RMA.

 

[Tahir2]

Try a different PSU even if its a crappy one just to see if the problem re-occurs. EDIT: Don't make it too crappy otherwise with that gfx card expect it to go pop!).

You got onboard gfx?

 

[Grall]

Don't have onboard GFX, but I have other GFX boards; I've switched my current ones out with one of my older 4890s and there's no difference. Heck, this PC ran with those boards initially and behaved just the same. I guess I would try with a set of NV G80 boards sitting in my old, unused Dell XPS box (which has a half-dead mobo (or PSU) too I might add, but that one has a whole different set of symptoms...), however I just don't think it would be a worthwile use of my time. Besides, I can't run with the G80s anyway, they're too slow in modern games. They can barely run Crysis at 1024*768 rez with the settings turned up, whereas my 6970s run it 60fps more or less @ full-HD rez even when downclocked to 500MHz core.

It's not the PSU either, because the crash only occurs unless I run the board hot first and then reboot it. If it was the PSU, rebooting would have no effect...

I'm gonna grab an old HDD some day now and do a clean Win7 re-install, and then do some diagnostic tests and see if anything interesting comes up. If not, I'll just turn the system in for warranty replacement. The external LCD Poster thingy's shot anyway, it only displays garbage if it displays anything at all, so I have reason to get the board replaced anyhow.

 

[entity279]

It's not the PSU either, because the crash only occurs unless I run the board hot first and then reboot it. If it was the PSU, rebooting would have no effect...

I wouldn't put my money on that.

 

[Grall]

Why do you say that?

 

[Mize]

I remember you mentioning this behavior when you first got your 69XX cards. It's very strange.
Two things stand out to me:

1. It doesn't error-out. It actually does manage to POST eventually rather than stopping and flashing and LED code-error on the motherboard (which would be quite helpful).

2. It has been progressive. This is strange since most motherboard components don't "age" over short time-spans. The one exception is capacitors. The large, cylindrical electrolytic capacitors near your CPU (and on other parts of the motherboard and in your PSU) age quickly compared to everything else.

If I had to guess, some power phase is on the ragged edge.

The fact that it's okay after a warm-boot points to capacitors also...they take a while to fully charge up and a long time to discharge.

What PSU do you have?

 

[Mize]

Grall:

What CPU are you running? At what clock?

 

[Tahir2]

Why do you say that?

Lovely post by Mize.

As to why it may be the PSU - unfortunately the only real way to verify if it is or is not the PSU is my suggestion to test with a different one. I had a second hand Seasonic recently that was suspect, the 12v rail was suspiciously low. It ran fine for two months, then would not allow me to reboot the system without discharging it for a few minutes and eventually stopped my system posting altogether.
I suspected the motherboard but had a spare PSU lying around from an el-cheapo old case, testing with that PSU resolved the issue. That is why I would point my finger first at the PSU but unless you test individual components then it is anyone's guess!

Bonus point: should be slightly easier than testing with a different motherboard.

 

[Mize]

Grall, were you able to give another PSU a try?
For reference I have twice had high-quality PSUs get to a "ragged edge" owing to capacitor aging. The symptoms were different each time, but the ones you describe fit the bill.

 

[Grall]

I remember you mentioning this behavior when you first got your 69XX cards. It's very strange.

It's actually always been this way, ever since I bought it. Well, the LED poster external dongle/thingy did work originally, and now doesn't really show anything anymore now.

1. It doesn't error-out. It actually does manage to POST eventually rather than stopping and flashing and LED code-error on the motherboard (which would be quite helpful).

Actually the bizarrely slow POST behavior stopped after those first couple boots. I'm not sure why. It's still not as fast as it should be (well, since it's an ASUS gaming board it isn't fast at all really), but it doesn't take minutes to get through the POST anymore.

If I had to guess, some power phase is on the ragged edge.

Maybe, but pushing the CPU doesn't crash the board. Only using the GPU(s) cause that to happen, and then only if the board hasn't been "run in" and rebooted, as mentioned... I don't even need to run anything particularly demanding, pretty much any game or GPGPU app using the GPU will cause it to happen. The windows aqua UI doesn't seem to cause crashes though; I guess it just isn't doing very much compared to games software. I've had a couple software crashes and BSODs in the past, particularly in IE9, but that may just as well have been a bad driver, it was never reliably repeatable.

The fact that it's okay after a warm-boot points to capacitors also...

Yeah, but why do I need to reboot the damn thing? Logically, it should be fine just by letting it warm up if it was a dodgy cap. It's like it's a combined software/hardware issue... *shrug*

What PSU do you have?

Enermax Revolution 85+, 1050W.

What CPU are you running? At what clock?

Core i7 920 (Nehalem, C0 spin.)

Stock clock and volts at the moment. 2.66GHz, but normally BCLK bumped to 166MHz (also unchanged volts), giving ~3.35GHz or whatever. RAM are OCZ Platinum 1600 DDR3 sticks, I first had 3, and then 6, running at either stock (1066MHz, yes?) or 1660MHz don't matter. Timings when tweaked, 7-7-7-20-1T, but stock timings and 2T command rate don't matter either.

Memtest x86 has run countless loops and no errors at any time, provided there's airflow to the DIMMs when running at 1660 speed.

Grall, were you able to give another PSU a try?

Sorry, I don't have any spares (that are ATX compliant, or at least that I dare use). There's an 1kW PSU in my Dell XPS box, but it has what appears to be custom plugs going to the mobo - thank you Dell... It actually has 2 main cable harnesses, one or both may fit in a standard ATX plug but god knows what'll happen to my hardware were I to try to plug either of them in... It also lacks the 8-pin ATX12V connector, so I'd have to juryrig something to cover that. And finally, it has only 6-pin PCIe power plugs for some stupid reason, so no modern high-end cards will work without faffing with adaptors and shit. Also I don't know if the wire gauge is dimensioned for 150W, they seem thick enough, but for all I know Dell put aluminium cables instead of copper, and they all turn into space heaters were I to plug in my 6970s... Gahh.

For reference I have twice had high-quality PSUs get to a "ragged edge" owing to capacitor aging. The symptoms were different each time, but the ones you describe fit the bill.

Yeah, I'm painfully aware of this possibility. The PSU was ordered from friggin' Germany (Enermax' retail presence is almost non-existant in Sweden), so that's where it'll have to go for warranty replacement.

Thanks all for the suggestions.

 

[Mize]

Enermaxes are good too...how old is it?

 

[Grall]

Same as rest of PC... Turns 3 in early 2012. 5 years warranty though! (On the PSU... Not all of the PC.)

 

[Davros]

Ive just repaired a pc that wouldnt boot
it was the mouse

 

[Grall]

Heh, my dad's HP portable is the same way. Won't POST if the dongle for his wireless mouse is attached. No frikkin idea what's going on there... It's just a pretty standard Logitech mouse, and the PC is not even a year old yet. Some weird UEFI bug I suppose. I told him to do a firmware update, but I know he'll never get around to it unless I do it for him.

 

[Davros]

I tried the mouse on my pc and it did boot but the mouse wasnt recognised

 

[Grall]

Davros, are you aware that someone or something (maybe related to the sock stealing goblins living behind every washing machine) has nicked all of your punctuation marks?