Sign in with
Sign up | Sign in
Your question

What could be causing a hard lockup?

Last response: in Motherboards
Share
March 13, 2010 2:32:20 AM

GA-EX58-UD3R, i7 920, Leadtek GTX295, 4x 2GB DDR2 Kingston Value Ram, 4x 1500GB Seagate HDD, Buffalo 1000W PSU.

My System freezes randomly mostly when playing 3D Games, sometimes if I'm doing nothing. But usually I can run it for several days with high Processor and Memory Usage with no problems. Seems like a Video Card Issue, but I switched it for an ATI HD 5570 and its exactly the same.

Certainly not a software problem because I can reproduce it on other OS like Ubuntu and Mandriva. I have fresh installed Windows Vista/7 several times, and it happens no matter which software or drivers I have installed. Pretty Sure it's not a Power Issue because I also tried switching that. My Hard Drive's have some problems but any 1 of them alone also triggers the freeze, so doubt it could be that.

I checked temperatures and they all seemed fine, the GPU often goes above 70, but the CPU is at most 50. Can't seem to get the other temperatures on the motherboard (if there is such a thing). When I take the case sides off and direct a large fan towards the motherboard it doesn't crash, but the temperatures are only a few degrees lower? I wonder if it is dust or something to do with the cables? After using the fan for a little while, I can turn it off and continue to play latest 3d Games, but after another day or two the problems will start again.

When the computer freezes it's completely unresponsive, so I have no idea how to debug. The Ubuntu and Windows logs don't seem to record anything. After a freeze, I have to reset and then it will freeze repeatedly at the BIOS Post (seems to be at the line Memory Frequency 1066Mhz, but sometimes elsewhere). Only if I power off for about 10 seconds and then turn on can I get back to normal.

Could be memory or motherboard or something I haven't thought of? I often run a few Virtual PCs in VirtualBox along with Vuze and other software, which uses a lot of Memory, and never experienced a problem. Even with 3D Games, the PC worked fine with that exact setup (-2 hard drives) for about 6 months, and just the last 3 months after installing I've put up with whatever the problem is.

I'm really just looking for some information on how to debug, or see if anybody has seen a similar issue. Could there be a BIOS setting I've messed up (though I don't remember changing anything except boot order)? If it's not the Video Card what could be causing 3d Games to freeze the entire system?

More about : causing hard lockup

a c 177 V Motherboard
March 13, 2010 3:45:26 PM

70°C+ is pretty high - could be thermal 'throttling' - TM1/TM2 throttling are detailed pretty well here:
http://ixbtlabs.com/articles2/intel-thermal-features/

You can disable TM2, but I'd not recommend it... Have you checked case air intakes/exhausts? 'Crud' collects incredibly rapidly - especially if you have pets :??: 
m
0
l
March 13, 2010 4:41:39 PM

Oops, I meant GPU temps at 70, edited now. I think thats still ok for the video card? Everest shows my CPU in 30s or 40s.

I am pretty sure it is a temperature or dust problem, because the big external fan helps me. But no idea how I could resolve this because its annoying keeping that fan there on full blast. I have one 9-10cm fan stuck on the back under the psu which seems to be blowing out, and one stuck between the hdd's at the front sucking air in (i think). To be honest, they hardly make any difference to temperatures. I have 2 other small fans but no idea what I could stick them to...?

I cleaned up as much dust as I could find, mostly on the CPU Fan and around the sides of the Video Card. Again, no difference, in fact my computer is reacting really strangely, lagging every few seconds. thought it was some mouse problem but sound also stops and starts. Still better than freezing but it could be a completely seperate problem. There seems to be constant 10-15% Hardware Interrupts in Process Explorer which I never noticed before. What on earth is going on?
:( 
m
0
l
Related resources
a c 177 V Motherboard
March 13, 2010 4:48:24 PM

GPUs are made to run hot, but the temps should not 'climb' over time...

How fast are you running the RAM?

Once again, on a hunch, download this and give it a try:
http://www.thesycon.de/deu/latency_check.shtml
Tell us what you find...
m
0
l
March 13, 2010 4:56:27 PM

Ok the machine stopped being laggy with all the Hardware interrupts after I closed Everest... so nevermind about that, though maybe its linked?

The Latency Checker is all in Green, Current Latency is around 80-120, Absolute Maximum is 4434.



m
0
l
March 13, 2010 4:58:49 PM

This is really interesting, it spiked into the thousands every couple of minutes.
m
0
l
a c 177 V Motherboard
March 13, 2010 5:15:48 PM

First thing to try is the latest BIOS release (& don't use @BIOS - it's a 'board killer')

...for rev 1.0 boards F11 here:
http://www.gigabyte.com.tw/Support/Motherboard/BIOS_Mod...

...for rev 1.6 boards FI here:
http://www.gigabyte.com.tw/Support/Motherboard/BIOS_Mod...

...for rev 1.6 boards, also FI here:
http://www.gigabyte.com.tw/Support/Motherboard/BIOS_Mod...

the last two might be the same BIOS, but hard to know short of d/ling & expanding 'em [:bilbat:2]



Comprehensive BIOS flashing instructions here:
http://forums.tweaktown.com/f69/bios-flashing-how-qflas...
m
0
l
March 13, 2010 5:51:19 PM

It seems if I disable the External USB Drive and Network Devices then I don't get anymore massive latencies every 2 minutes. Is this good or bad? I don't see how it could be a device driver issue though because it can even freeze before any OS is loaded.

Was a bit paranoid it would freeze whilst installing the FI BIOS upgrade with Q-flash, but that went OK. So far no crashes, but I'll leave it downloading overnight and see how it goes. Goodnight. :) 
m
0
l
a c 177 V Motherboard
March 13, 2010 6:36:49 PM

Give 'er a try - have some alternative things to check if the new BIOS doesn't do it for 'ya!
m
0
l
March 14, 2010 3:31:20 AM

Yea, it didn't freeze overnight, but did after playing a random 3D game for a few minutes, just as before. Even when I disabled all those devices which caused the spikes it still froze.

Really wish there was a way to detect the exact event or DPC or whatever which caused the lockup.

:??: 
m
0
l
a c 177 V Motherboard
March 14, 2010 1:06:19 PM

Quote:
Really wish there was a way to detect the exact event or DPC or whatever which caused the lockup.

:lol:  ...pretty much, everyone who's ever had this problem wishes so, too!! :lol: 

There are a couple places to start looking, however; and a couple techniques to use to look...


First, I have a question - did the BIOS replacement seem to 'cure' the pre-OS problems? In other words, are you still getting 'freezes' prior to loading up an OS, or always after? If after, are the problems still 'reproduceable' in differing OSs?
m
0
l
March 15, 2010 6:01:37 AM

Unfortunately it's exactly the same, Mandriva is the worst, freezing after just a couple of minutes without doing anything. Ubuntu crashed after I ran Modern Warfare 2 through Wine. Pretty much exactly the same as if I was running it in Windows, lasts maybe one hour on lower settings, a few minutes on higher settings. Looping audio and no response from the keyboard.

On Ubuntu there is another kind of freeze where I can still move the mouse, probably unrelated.

I tried many different games on Windows, all the same, except a few older 2D games. If I keep a fan on, even on lowest power, then no problems.

After any freeze, if I reset it gets stuck in the BIOS after just a few seconds. If I press reset again and again it doesn't get any better, though sometimes freezes at different points during the post and boot, even whilst windows is loading. My instinct is that some component does not completely reset itself unless it loses power for a few seconds?
m
0
l
a c 177 V Motherboard
March 15, 2010 2:03:26 PM

It's beginning to sound like you might just need to 'stir things up a bit', and just get some better airflow through the case; even for non-gaming machines that I build for other folks, I automatically stick in a pair of 120mm PWM fans (one intake, one exhaust), so I can 'daisy-chain' em off the board controller - hardware cool = fans quiet; hardware hot = fans blasting... If the case only will take 80s, I use a pair of intakes, and a pair of exhausts. My workstation is water-cooled, and still has six 120 case fans! The problem is: it doesn't matter how good your CPU sink is, or how well 'aerated' the GPU shroud is - if you'r just circulating the same hot air through 'em, over and over again, the heat just has nowhere to go :cry: 
m
0
l
March 16, 2010 3:01:00 AM

Thanks for your advice, I really want to put more fans in, but it's just too cramped already. There doesn't seem to be any area where I can stick a single 120mm, let alone six of them. Shouldn't there be holes in the casing for that? Do you mean fixing smaller ones one on top of the other? I'd like to buy a water-cooling system or something like that but just need to rule out a broken motherboard. The machine can even crash with front, sides and top of the case removed, where Everest doesn't show any temperatures over 40 degrees... and here in Beijing it's like snowing every day... so I thought there might be some faulty chip on the motherboard which is prone to overheating (and which hasn't got a sensor). Or worse, the sensors are just broken?
m
0
l
a c 177 V Motherboard
March 16, 2010 2:02:37 PM

Beijing?? Cool! Never have been to China - and always wanted to go - now more than ever; trying to learn Manadarin, having gotten interested through the calligraphy - have probably inked 'yŏng' ('forever') ten-thousand times! Diplomatic corps? Or just incredible well-educated? Your English is impeccable...
Quote:
Shouldn't there be holes in the casing for that?

Yes indeed - it's very difficult to make a hole big enough for a fan - I only buy cases that have sufficient holes from the manufacturer. I'm pretty good with a 'nibbling' tool, but by the time you can use one to make an eighty mm hole, much less a one-twenty, your wrist will have fallen off! :cry:  I think I mentioned, my criteria is: either one hole in the front and the rear (or in the top panel, near the back) for 120mm fans, or a pair of holes, both front and rear, for a total of four 80mm fans. Another thing 'enough' fans gives you is the ability to run 'em a bit slower, while still moving enough air to keep things reasonably cool...

Thought a bit, and have another thing you might try - if you haven't already... Try reseating the video card, and its power connectors. My thinking is: maybe (long shot...) the connector's contact 'fingers' are expending from heat, and breaking connection a tiny bit.
m
0
l
March 17, 2010 10:59:37 AM

Hehe, I'm just a British person living here. I'm studying machine translation in a Chinese university, so I have to be able to speak and read Chinese to get through classes, but to be honest, writing by hand and calligraphy are things I find extremely hard.

It sucks trying to build and maintain your own desktop here. Half of the components they sell are fake in some way and it's impossible to return them or get a replacement. And I find it very hard to trust the PC repair shops. :( 

My case has an 80mm fan area on the back but that's it. I could fit another 80mm fan under the graphics card by taking those expansion slot cover things off, but no idea what to do about the front... ? May be worth getting another case, but thats almost as much trouble as replacing the motherboard... :??: 

I've changed the video card a few times and no problems, can't see anything wrong with the connections... I realised I'd never tried using the other PCI-E slot, and just gave that a go, but my computer completely refuses to get past the post (frozen at "Memory Frequency..." just as if I had reset it after a freeze). Maybe this can narrow down the problem a bit? In the process of moving the card back I managed to burn my finger on this grey heatsink thing to the right of the PCI-E slots. Is that meant to be so hot? Cheers.
m
0
l
a c 177 V Motherboard
March 17, 2010 2:01:12 PM

The calligraphy is difficult, but, once pursued, beautiful, and very relaxing. I imagine there you do it some other way, but the 'ancient art' gives you an 'ink stone', with a 'well' carved in it (and, if you can afford it, you can get a millenia or two old one, with a dragon or an egret carved in it, as well...), into which you drop water a drop at a time from a tiny painted ceramic dispenser, and 'rub' an ink stick (that might also be six hundred years old, pressed into the shape of a dragon, or a Buddha...) around in it, for a half hour or fourty minutes while you 'meditate upon' your intended strokes, characters, and message... Infinitely calming - I think it drops my blood pressure about twenty points every time I practice :D  I think it took me, like, a year of practice to get the 'nà', the 'falling right hand' stroke, as it requires that you 'dip into' it toward the end, and then smoothly withdraw the brush, to get a diminshing 'tail'... One thing I can't recommend highly enough is this book:
http://www.amazon.com/exec/obidos/ASIN/0966075005/zipub...
and the same guy's website, here:
http://www.zhongwen.com/
The book has a system for 'classifying' characters,that makes it actually possible to find an unfamiliar one (well, at least, most of the time [:bilbat:3] )

The 'gray thing' is the heatsink on your southbridge, the Intel ICH10R that 'runs' your drives, the two PCIeX1 slots, the lan chip, and the USBs. While these are made to 'run hot' (Tcase-max is, at its lowest, 105°C), it seems odd to me that it would... The VidCard is actually 'hung off' the northbridge, the X58 IOH. Tell me about what all else is 'hooked up' - any other cards besides the VidCard? How many drives?
m
0
l
March 18, 2010 3:36:18 AM

I have that book, its very good. Essential when I was learning Chinese back in England (translating all those poems etc :??:  ), but nowadays all the chinese I read is in electronic form, and got in a somewhat bad habit of relying on browser plugins for pop-up pinyin. But I still use it every now and then of course. My trouble is that I can recognise thousands of characters by now, but could probably only recall how to write about 20% of them. That book is cool because it's full of very interesting facts and links between characters, which Chinese people just never heard of before... naturally, Chinese people I met tend to view the book as a load of rubbish from foreigner's imagination...

The southbridge is attached to just the one PCIeX1 card, no other add-on cards. 4 SATA hard drives and a SATA DVD-RW... don't know if they're relevant? I have a lot of random USB things connected, but the problem still occurs even if I disconnect all the USB and SATA.

Usually the BIOS post will read...

...(Few lines of Motherboard and Processor specifics)...
Memory Frequency: 1066 Mhz

Detecting IDE Drives...
[List of those 5 SATA Drives]

etc.

But all the freezes seem to make it crash before displaying "Detecting IDE Drives...", which usually takes less than a second to come up. So my amateur opinion is that the overheating thing (southbridge?) is involved at that stage of the post, just before it checks for drives.

About Calligraphy, many westerners I've met here in China tell me that Calligraphy helps you to understand every facet of Chinese culture, and to be geuinely accepted and feel a part of Chinese society. I think this is quite fair, so I've always wanted to give it a try, but I'm such a clumsy and typically unartistic and unspiritual type of person that I'm constantly aware that I wouldn't get very far. A little bit sad perhaps :) 

m
0
l
March 18, 2010 3:48:15 AM

It seems the southbridge doesn't have a temperature sensor, but when I touched it, could certainly be close to 100 degrees.
m
0
l
a c 177 V Motherboard
March 18, 2010 2:19:01 PM

Quote:
But all the freezes seem to make it crash before displaying "Detecting IDE Drives...", which usually takes less than a second to come up. So my amateur opinion is that the overheating thing (southbridge?) is involved at that stage of the post, just before it checks for drives.

Yup - that sure does make it sound like it's some kind of SB problem...

One possible solution is to put fans right on the bridges:



There is a problem, however (beside the fact that finding the little fans there may turn out to be a PITA :cry:  !) - GB, for reasons known only to them, seems to, on a lot of boards, 'impede' the cooling fins' operation by putting little 'bling' covers over the sinks! To put a fan on, you pretty much have to 'pry' the cover off, which, I'm guessing, voids the warranty :(  Should you wind up needing to RMA the thing, you'll likely be SOL!

If you can get the fans, and want to give it a try, I can 'hunt up' a post I did six or eight months ago, that goes into detail about how to do the project...


Quote:
Chinese people I met tend to view the book as a load of rubbish from foreigner's imagination...

That's very amusing - I would have never guessed that - it all seems so, well, coherent! Does seem though, to help me remember characters, at least a little [:bilbat:2]
Quote:
I'm such a clumsy and typically unartistic and unspiritual type of person that I'm constantly aware that I wouldn't get very far.

:lol:  I, on the other hand, am kind of an 'artsy, craftsy' type - here are some pieces of my pottery:


Oh, and can't forget this little guy (roughly three inches tall...):



m
0
l
March 18, 2010 3:51:15 PM

Hmm, imageshack seems to be blocked in China, but I got to see your pictures through a proxy, very interesting, especially the alien and it's teeth. Must take a huge amount of patience as well as skill... Why did you write 傻子 on the flowers by the way?

I think from "fan off => freezing, fan on => no freezing", we can be sure that it's an overheating problem. Because the temperatures don't seem insanely high, it must be some area which hasn't got a sensor attached, or the motherboard is broken in an obscure way.

There's not much chance of me RMAing the motherboard from here, but I doubt I'll be able to get a hold of those special fans either. Probably best if I just keep the status quo (turning on the fan when I want to play a 3D game) and save up for a new motherboard and case. When I look back, it was kind of dumb of me to go for the case which looked the nicest. It's so cramped and there's nowhere to put fans. Envy so much the person who has the workstation in the photo... wiring and everything so incredibly tidy. But why a fan on the base? Watercooling looks less complicated than I imagined though that may be an illusion due to the insane tidiness.

I investigated the bridges a little. The seem really loose, you can almost pull them off, is that normal? Think I also see what you mean about the screws.
m
0
l
a c 177 V Motherboard
March 18, 2010 4:51:29 PM

Quote:
Why did you write 傻子 on the flowers by the way?

:lol:  gotta figure that out myself! It's one of the very few I've done that is not engraved on the bottom with the translation... The other says "All kinds of strange things overwhelm the fool; the sage pays close attention to his own train of thought" I almost think the first was meant to say "This space left intentionally blank" :pt1cable: 
Quote:
When I look back, it was kind of dumb of me to go for the case which looked the nicest. It's so cramped and there's nowhere to put fans. Envy so much the person who has the workstation in the photo... wiring and everything so incredibly tidy.

it's mine - and I'm an obsessive compulsive! :pt1cable:  Manic-depressive - I actually taught myself calligraphy, and learned ceramics during a stint in prison; mania makes you insomniac, and fills your head, pretty much with miscellaneous crap running 'round in ever-tightening circles; alcohol (and/or drugs) change that (you think, at the time, for the better, especially when you finally pass out) but form a feedback loop: first the stressor is the mania, which filling your brain with dopamine from booze/drugs alleviates - pretty soon the stressor becomes 'lack of dopamine', by which time you're pretty thoroughly screwed! Wasn't diagnosed 'till my mid-to-late fourties, and took a decade of trying various and sundry combinations of psychiatric meds to find something that would work for me, without poisoning my liver and kidneys, and without turning me into a zombie (which, of course, would lead to me going off the meds :??:  ), as well as finally 'stumbling over' some actually effective therapy (cognitive/behavioral)!

Had to learn a couple things along the way - the first, and most important of which was "passing out is NOT sleeping - and, conveys NONE of the benefits, to your mind and well-being, that actual sleep does! [:bilbat:7] Oh - and don't worry - I'm not an axe-murderer or anything - just drunk driving (once too often :cry:  )

The box is a CoolerMaster Cosmos - which I can highly recommend - if I'd empty out the computer junk, I could likely park a small car in there... Even though it's water cooled (just the CPU and VidCards), it has a half-dozen 'case' fans; three exhaust (two in the 'roof', one at the rear); and three intakes, the visible one at the base, which came with the case, and it uses a sort of 'tight' drive rack, which is full, so I added a fan at the top and rear of the rack to 'suck' air through the drives. The advantage of lots of fans, is that I can run 'em all pretty 'slow and quiet', and still get excellent airflow, and cooling...
Quote:
I investigated the bridges a little. The seem really loose, you can almost pull them off, is that normal?

The sinks are likely held into the bridges by a little 'push-through' plastic locking pin, which has a spring to tension 'em against a thermal transfer pad. The pads themselves are not all that efficient at moving heat; I yank 'em, pull the pads, lap 'em (which is mostly just obsessive-compulsive :pt1cable:  ), and use thermal compound in between. The trick, though, is afterward, you need to replace the pins with tiny (4-40, I think) nylon screws and nuts, as the 'stack height' gets lower without the pads, and the stock retainers won't produce enough pressure through the springs to hold 'em in contact properly...
m
0
l
!