Intermittent lockup on XP, trouble shooting ideas?

[delta_charlie]

Hi all, I have a HTPC that has developed an intermittent lockup problem. Most of the time it locks up it is running ProgDVB but some times it locks up when running Winamp with streaming audio. When it does lockup it does it bad as the only thing that clears it is a power off. The locking up does seem to be related to how hard the computer is working as it locks up more often when running ProgDVB but I do not believe it is the software that is causing the trouble as a (CTRL-ALT-DEL) does not do anything.

The motherboard was recently replaced (MSI P6NGM) under warranty because the first one died and would not post.

I tried reloading the sound, audio and video drivers and it seemed to help some but the locking up continues.

I can monitor the CPU temp from the BIOS setup screen and it runs cool.

No errors or any hint of what might be wrong. In an effort to try and find out what might be causing the trouble - I ran it all day with Linux and had no problems. I would really like to try and figure out if it is a hardware or software problem.

Anybody think reloading XP would help?

Any ideas?

Thanks, DC

 

[8bitbytes]

Any errors recorded in the Application and System Event Logs?

 

[gbjbany]

It could be memory, it could be you are only finding this when certain memory is accessed.
You could try running one of the memory test routines that will let you know if your memory is ok. but here's thread with a few links to tools Memory stress/burn-in programs - Overclockers Forums
http://www.ocforums.com/showthread.php?t=123908

I assume you have checked it is seated correctly,

 

[diogen]

I can monitor the CPU temp from the BIOS setup screen and it runs cool.

When rebooting? That's too late to catch CPU overheating.
What are you watching using ProgDVD and what codecs do you use? What is your CPU?
Pushing CPU utilization to 100% when watching hidef (especially interlaced) is a piece of cake...

Anybody think reloading XP would help?

You haven't done this after replacing the motherboard? Then you should do it.

My guess would be it is video card/driver or heat issues (CPU or GPU).
RAM problems most of the time manifest themselves in reboots...

Diogen.

 

[mike123abc]

My experience is that hard lockups are usually drivers. Memory or other hardware issues usually seem to give the blue screen of death or a reboot. I would check that you had the latest BIOS, and drivers for all the hardware.

 

[delta_charlie]

Hi all, thanks for the replies.

From 8bitbytes: "Any errors recorded in the Application and System Event Logs?"

I just checked the system event logs but did not find anything near the time the computer last locked up.

From gbjbany: " It could be memory, it could be you are only finding this when certain memory is accessed." Good point as I believe the computer uses shared memory with the video. I will try running the test you linked to.

From diogen: " What are you watching using ProgDVD and what codecs do you use? What is your CPU?
Pushing CPU utilization to 100% when watching hidef "

I'm just watching SD, using FDShow codecs. The CPU is a dual core Intel 1.6 GHz and it is only running about 15 to 30 % CPU utilization.

From mike123abc: "My experience is that hard lockups are usually drivers. Memory or other hardware issues usually seem to give the blue screen of death or a reboot. I would check that you had the latest BIOS, and drivers for all the hardware."

My gut feeling is it is a driver or video problem. The computer has no other problems and boots fine. Perhaps it is just the video locking up. The thought just came to mind that maybe just the video is locking up but the computer may still be running. I just tried the power off key but for some reason it does not work.

I would like to try a key press shutdown the next time it locks up test to see if the computer is still running. Is there a key press I could use to try and shut down the computer?

Thanks, DC

 

[8bitbytes]

Once your computer locks up, the only way to do anything is to hold the power button for 10 seconds to kill the power or pull the power cord.

Because there are no error entries and it locks up when pushing the cpu, You are most likely overheating the cpu. It might be a bad ram stick since you are also filling up more ram running these apps but I would suspect cpu overheat first.

 

[delta_charlie]

Update - I'm still working on it :-(

Dang! this sure has been one tough problem to find. So far I have re-installed XP Pro, stopped the updates at SP2 to test. Installed all the latest drivers I could find for the video, audio and mother board. Tried Mytheatre, and the latest ProgDVB and the intermittent lockups continue. This morning I re-installed GRUB (the computer is a dual boot XP and Ubuntu box) and got Linux back up after the XP re-install. Just noticed this morning that Ubuntu comes with memtest86+ and can be started at the boot prompt so I running it now and it has made 3 passes with 0 errors.

Will try running memtest86 again in the morning when the computer is cold. The lockups seem to come most often about a hour after a cold start.

Maybe it's about time to take the ax to it, that might make a good video for youtube ;-)

Will post follow up if I ever figure out what is wrong with the darn thing.

Later, DC

 

[Van]

A couple of things to try.

Disable all unnecessary programs and see if it continues, if not then you can start by disabling one at a time till you find the one that's causing it and replace it with an alternative program.

Replace the programs that your using now for video and audio playback and see if it continues.

Try removing a stick of ram and see if it continues,

Open the case while the programs are running and make sure the cpu and video fans are running and make sure the fans and heat sinks arent clogged.

Take a look at your programs and see if any are old 16bit programs, lockups are common with them from what I remember reading a while ago and could be whats causing yours.

 

[delta_charlie]

Think I'm closing in on the problem

Hi all, I think I have the problem narrowed down to the MB or the CPU or maybe the power supply.

This morning when I ran memtest86+ from a cold computer about 40 minutes after starting the memtest program locked up. This narrows down what could be causing the trouble as memtest86 runs from a small Linux OS bypassing all of the windows XP.

I found it interesting the cursor was still blinking and so was the + sign.

What makes the cursor or the + blink? Is is being controlled by the CPU indicating that the CPU is still running?

Anyway to isolate the remaining components to see what is causing the trouble?

Thanks, DC

PS - If I was to start on this problem again the very first thing I would do now is run the memtest86. I could have saved a lot of time as it looks like the problem has nothing to do with XP or any of it's software or drivers.

 

[diogen]

Take off all heatsinks (CPU, GPU, NB/SB, etc.), clean them and the chips, reapply Arctic Silver 5...
Leave the case open during testing.

Diogen.

 

[snathanb]

What makes the cursor or the + blink? Is is being controlled by the CPU indicating that the CPU is still running?

The blinking is controlled by an ASCII code, not the CPU. Once that code has been sent, it will continue to blink without CPU intervention.

I am going to second those who said memory issues. I've had this problem with some cheap RAM in the past.

 

[Agonizing Fury]

I had a similar problem a few years ago with a computer I built. It would only lock up when I was running the computer hard. Then I started to have hard drive issues as well (which I originally attributed to all the forced reboots). It wasn't until I started to monitor the +12V power that I found out that my +12V would drop as low as 9 volts during heavy CPU usage. So, I replaced the power supply (whith a much beefier one I might add). Same problem still. Problem wasn't fixed until I replaced the motherboard. Apparently something was wrong with the MB that was causing the voltage to drop very significantly during heavy CPU usage. I still to this day don't what caused it to happen.

 

[delta_charlie]

Another update

I called tech at MSI and they suggested I try erasing the CMOS so I did last night but it did not help the problem.

I also loaded the dual core software that lets me monitor the CPU temp and voltages while running the computer. The CPU runs 33c when running the ProdDVB and all the voltages on dead on with no variations when the HD is running or not. The only thing I noticed was the DRAM was running at 667MHz but I thought my memory was 533MHz. Did not see a way to change it from the BIOS. Not sure what is up with this; however, I have some new 667MHz memory due in today so I may try it. I also ordered a new ASUS MB and new 800MHz memory to go with it. I might just pull the MSI MB and replace it with the ASUS.

Got to run, DC

 

[Van]

Is the bios current and is it the correct bios or a modded bios from a modding site? Pull the memory sticks and check the contact points of the sticks and down in the module slots, look for anything that seems out of whack.

 

[delta_charlie]

Looks like it may be the memory

Hi all, the new Kingston 667 MHz memory came in yesterday and I replaced the old memory last night then ran the memtest86+ program and got two good passes on it so I decided it was time to play and rebooted into XP.

ProgDVB has been working great with no lockups so far. I might still have trouble but it is now working better then it ever did with the replacement MB.

Speaking of the replacement MB, when I had the MSI tech on the phone the other day he said they did not actually replace the MB but repaired it instead. I'm thinking whatever they replaced on the MB may have changed the supported memory as the computer used to work just fine with the old memory before the MB broke and would not POST. I believe the old memory was suppose to run at 533MHz but I noticed the (fixed) MB had the DRAM set to 667MHz

In any case the new 667 MHz memory seems to have fixed the problem.

It has been both frustrating and an interesting learning experience trying to fix the computer. I learned a lot and I'm glad I did not take it to a repair center as I would not have learned near as much. Now if I have any computer troubles like this again I will know to check the memory first.

Got to run, DC

 

[8bitbytes]

Thanks for keeping us informed! Glad to hear things are working again.

 

[TheForce]

I had issues with not having enough ram to work on this computer. It was a single 512Mb stick. In the past all worked fine but over the years, this computer (P4 3.6 Ghz. ) was being used with Firefox and multiple tabs. It just got too slow but otherwise ran solid. So I pulled the 512 out and replaced with 2- 1Gb sticks and while everything ran faster, especially Firefox, it would often just black out and reboot. I took the ram back and exchanged it and had similar problems. Finally, I returned the ram and replaced with a single 2Gb stick and installed it in the same slot as the older 512 Mb originally. It has been running fine ever since and rarely needs reboot due to memory.

I think faster computers are more sensitive to matching memory modules.