Plagued by strange system problems

[ow_tiobe_sb]

Of late I have been struggling with an irritating series of occurrences affecting the stability of my PC.  Here is some vital information about my system:

OS: Windows XP Professional running SP2 (up-to-date with all patches and security gizmos)
CPU: Intel Pentium 4 3.06GHz
RAM: 1 GB DDR PC2700 RAM (2 512 MB sticks in a total of 2 available slots) and 256 MB nVidia GeForce 6600 AGP video card (drivers up-to-date)
Fixed Drive Capacity: C: 160 GB (16 GB free space); F: 120 GB (12 GB free space)
Anti-Virus and Anti-Spyware Software: Norton AntiVirus 2005 (up-to-date) and Webroot SpySweeper (up-to-date)

The Problem: Intermittent system freeze, during which various programmes (e.g., Photoshop, 3ds max, Firefox, explorer.exe) hog 99% of the CPU usage.  This seems to happen without warning and does not seem to be attached to any system event, such as application close or open. 

I have noticed that, on a number of occasions, the following will happen:

1. A programme, typically (but not always) memory-intensive, will cause the system to freeze due to 99% CPU usage.

2. If I am lucky, I will be able to CTRL-ALT-DEL, pull up the Task Manager, and kill the CPU hogging process (often there is a long wait due to overloaded CPU).

3. What typically happens next is that explorer.exe will shoot to the top of the CPU usage hill, taking over that 99% usage and re-creating the scenario in Step 1.

4. If my luck continues, I will be able to kill explorer.exe and log off and log back in.  This will fix the problem for a time until it starts all over again.

5. If I am not lucky, I may be forced to turn off the machine physically without shutting down, restarting, or logging off.

6. Occasionally, if I am simply trying to shut down whilst the 99% CPU usage problem exists (and even when it is not present), the computer will automatically restart.  I have already unticked "Automatic Restart on System Failure" in the System Control Panel, so I have no idea why this should be happening or if it is directly related to the CPU usage problem.

Other Potentially Useful Facts:

1. I have done complete virus and spyware scans of both drives.  Nothing is coming up.

2. I have done complete disk surface scans of both drives.  No bad sectors (and I understand that this can be misleading on modern hard drives due to reallocation) or other problems of note.  I have also run Norton Disk Doctor on both hard drives.  No problems to report.

3. I invested in a programme named "AntiCrash" to help prevent unseen problems from freezing up the system.  To date, it has not prevented this problem from occurring.

4. I have also installed HDD Health, which uses S.M.A.R.T.  It has noted a high temperature of 50 degrees F on the slave drive, Drive F: (since I have removed the drive from the tower, it runs at anywhere between 30 to 38 degrees F), and has logged raw read errors and reallocated sectors (I do not have the exact values available at the moment.).

I am hoping against hope that I am not staring down the barrel of an imminent hard drive failure, for I do not know how best to back up that amount of data on a slave drive (I do not think I have additional data and power cords available on the motherboard to hook up a third hard disk for data transfer.  Perhaps booting from the DVD/CD-ROM drive would allow me to remove the good drive containing Windows, install the destination drive, and copy the data over to the new drive?).  Moreover, I am not certain that the system would remain stable enough to make such a transfer possible without freezing up.

If any technical gurus have any advice, I would appreciate it.  Thanks, in advance, for your responses.

ow_tiobe_sb
Phantom Bunburyist and The Prat in the Hat

[BlueBard]

One thing you might try, if you haven't already, is to defragment your hard drive(s).  Windows seems to do squirrely things when the system drive is fragmented.

Obtain and install a Wininternals utility called PageDefrag (available from Microsoft's website, who now own the Sysinternals folks, PM me if you need me to find the link for you.)  This utility will defragment your registry files and the system page file which the normal defragmentation tools can't touch.

If that fails to improve things, I might be able to suggest some more freeware utilities that could help.

[Dweomer Knight]

As with most things related to computers, there are a number of things it could be.

The first thing I would check (if you haven't already), is your computer's Virtual Memory settings.  Right-click on My Computer.  Select Properties.  Click on the Advanced tab.  Click the Settings button under Performance.  Click on the Advanced Tab.  At the bottom you should see how much memory is dedicated to Virtual Memory (which is basically harddrive space that the computer uses as pseudo-RAM).  If this setting is less than 1000 MB, switch it to at least that high.  On a side note, I would upgrade to 2 GB of RAM as soon as possible.  When I did that to my machine I noticed significant improvement across the board.

The symptoms you describe I've never seen as an indication of HD failure but that doesn't mean it isn't possible.

There could also be a corruption in Windows XP itself or one of it's plug-ins.  Those can be tricky to track down though.  One option is to go to Microsoft's website and post your problem there.  A lot of tech wizards read those and can often help with the more obscure fixes that auto and even high-priority updates don't cover.

That's all I've got off the top of my head.  I'm sure others have some additonal ideas as well.

dk

[ow_tiobe_sb]

Thanks for the quick reply and advice, BlueBard.  I have downloaded PageDefrag and will give it a go.  I forgot to mention that I use "Registry Booster" to scan, clean, and defrag the Windows registry regularly, so I do not think the problem lies within the registry.  It remains to be seen whether or not the page file is to blame.

I should also mention that I ran System File Checker ("sfc /scannow" at the command prompt) last night and received no results whatsoever.  The checker simply quit after completing its survey.  Is this normal?  :unsure:

EDIT: Thanks for your reply, DK.  Recently raised the Virtual Memory from somewhere in the vicinity of 1600 MB to 2500 MB combined over both hard drives, and one would hope that would be sufficient.  This amount exceeds the amount recommended by Windows, so I am wondering if I should let the system manage my Virtual Memory...

ow_tiobe_sb
Phantom Bunburyist and The Prat in the Hat

[BlueBard]

Oh, and backup is essential.  Anything that you can't recover by reinstalling from the factory CD(s) needs to be copied somewhere, somehow.

You might pick up an external USB hard drive and back up to that.  If you can't bear that expense, you might try buying Norton Ghost.  It will allow you to image your hard drive in a number of ways to various destinations (CD, DVD, USB, Network, etc).  OR, if you have write access to an FTP server somewhere, you can use a free program called G4L that boots from a CD and will image a drive via FTP.

[BlueBard]

Ordinarily you want Windows to manage its' own Virtual Memory.  If you do that and you get tons of messages complaining that it is low on virtual memory and needs to extend it, that's a good sign that you need more RAM.  Generally though I'd think you'd be good with 1 GB RAM.

You might also want to check and make sure that you don't have too many programs starting up at boot time.  There are freeware utilities out there that can help you with that. (I have a feeling you may already have one!  :lol:)

[Symon]

Hi, ots.

Firstly, I WOULDN'T, under any circumstances let windows manage it's own virtual memory. Windows couldn't manage it's own memory if Billy's life depended on it. Amongst other things, the way it dynamically grows the amount of swap it guesses it needs means it is fragmented. Fragmented swap can be worse than insufficient. (You don't mention your hard disk thrashing when CPU usage hits 99%. I've seen windows dynamically expanding/using fragmented swap produce similar symptoms, but the hard disks thrash like fury! Still, with both disks so full, windows will struggle allocating swap, and it sure won't be contiguous!) Nor would I pay any attention to its recommendations for virtual memory. Have at least twice as much swap as you have real memory, unless some software you use recommends more. I don't usually go for more than 2.5x. Best way to set it up is to turn off all swap, reboot, defragment, reboot, set up manual swap. Other than that difference in professional opinion, you've had good advice so far.

Now then, possible sources of the problem. Is your processor overheating slightly? (What does your bios claim? Does this occur after a long session?) Are your memory timings off? I've seen this produce similar effects (Does your bios allow you to set them? Check the motherboard manual if you have one.)

Your F: drive does sound a little suspect, but if at fault, this shouldn't happen if you don't access that drive. One way to test it is to disconnect that drive and see if you still get these symptoms.

Good luck, l'll do what I can. (As you know I do this sort of stuff for a living.)

[catwhowalksbyhimself]

Actually, I have another possibility.

Do you happen to have more than one RAM chip?

I ask this because anything that plugs in can wiggle out of it's plug and a RAM chip not quite fitting in quite could spontaneously reduce the RAM, causing programs that don't normally take much memory to suddenly take all of it.

I have a feeling I'm on the wrong track though. . .

[BlueBard]

(As you know I do this sort of stuff for a living.)

Heh.  I used to.  But I've spent more time dinking around with networks, Linux and such as of the last few years and I'm no doubt rusty.

The ONLY reason I suggest letting Windows manage its' own virtual memory is because IF you have enough RAM and enough free disk it leaves you alone.  It's too easy to forget that you've set it manually when there's an issue and most users won't even think about it.  But, what he said, OTS.  Unplugging the suspect drive (since it isn't a system drive) is a good suggestion.

And cat's suggestion isn't a bad one, either.  Reseating cable connections, RAM modules, and boards can make some odd behaviors go away.  A problem I was working on for someone else recently turned out to be caused by a loose screw.  (In the computer, not the ones in my head.)

[stumpy]

Cat, I'd say your instinct is good that this is a memory swapping problem, but I would be very surprised that the RAM is coming loose without near-constant BSoD crashes. At least unless home PC motherboards have gotten much more sophisticated about hot-swapping RAM since I bought one. Still possible, though...

But the overall symptoms of freezes when high memory applications are running on a system with a short supply of likely fragmented swap space probably points toward what people are suggesting - a very busy page file. Throw into the mix the possible issue with the F: drive and the fact that the swap space also partially resides there and that seems like the likely bet.

ow_tiobe_sb, I would use the method described to really defragment C: (and maybe F:) and then set the virtual memory to twice real RAM, all on the C: drive. See if that clears things up.

In the meanwhile, I am liking USB/1394 drives for back-ups and, with your F: possibly looking to take a headlong plunge in the the great celestial bit bucket, you want to have backups. It will make recovery much less painful. If USB drives (probably a $100 model would allow you to copy your F: drive and 500 GB models are under $200) are too dear, but you can afford another internal drive, then you can probably do what you need for backups with the cabling you already have. Most PCs have two EIDE (or ATA) channels, each with a master-slave cable. You are using one for the two hard drives you have and the other is likely running your CD/DVD drive. That second cable should have an available connector for a third drive, or you could make one available by swapping out your CD/DVD or whatever other storage device is taking up the extra connector.

[ow_tiobe_sb]

I am truly grateful for this tremendously helpful response to my request for assistance.  Many thanks for your knowledge and advice.  I was wondering if anyone could recommend reliable and affordable brands of external USB hard drives.  I used another free application last night, SpeedFan, to detect the temperature of my CPU.  It also told me more by way of interpreting hard drive S.M.A.R.T., and it confirmed that my F: drive is indeed in a bad way. :(  Consequently, I will want to address the backup issue quickly.

My current theory (which I have culled from the theories of others addressing other computer problems posted on the Web) is that the problems with the F: drive are making the CPU work harder and causing it to overheat.  I can part with a drive, but I would rather not part with my CPU. :ph34r:

Symon, a few words leapt from your mouth and stumbled into my tin ear for technology terminology.  Could you define "memory timings" for a layman?

Thanks, again, for all of your very good counsel. :)

ow_tiobe_sb
Phantom Bunburyist and The Prat in the Hat

[ow_tiobe_sb]

My problems persist.  Here is an update:

1. I purchased a 250 GB USB drive and completely backed up the F: drive (with the exception of a few files--sadly, FF-related files :(--that the drive could not access well enough to copy).

2. I have removed the F: drive completely and reassigned the F: drive letter to the USB drive so that all programs installed on that drive continue to operate smoothly.

3. I have uninstalled useless programs and freed up a great deal of space on the C: drive.

4. The C: drive has been completely defragmented and the pagefile has be defragmented.

5. I have removed and reseated both RAM DIMMs.

6. I have reconnected all wires.

Nevertheless, my system freeze problems still recur sporadically, and the strange auto-restart problem persists (and this is perhaps the most regular and most annoying problem, for, at times, I cannot seem to turn off the machine without removing the power cord from the rear of the tower).  I have the auto-restart function deactivated in the Windows XP settings, yet this problem remains.  Any ideas?  :unsure:

ow_tiobe_sb
Phantom Bunburyist and The Prat in the Hat

P.S. I have installed a programme called SpeedFan to detect my CPU temperature and fan speeds, but I am having some trouble interpreting the findings.  If I am reading it correctly, it seems to indicate that the fan speed rate is 44%.  With this programme I can manually increase or decrease the fan speed rate, but setting it to 100% to try to decrease the CPU temperature does not seem to make a difference in system performance.  I will try to take a screen capture of the programme's readout to share it with those of you who might know better how to interpret these results.

[Panther_Gunn]

at times, I cannot seem to turn off the machine without removing the power cord from the rear of the tower

I'm not sure how common the awareness of this is, but most computers nowadays will force a shutdown if you hold the power button in for something like 5 seconds or so.

The only other likely candidates that come to mind for the persistant problem would be a failing C: drive (way back when, I had one that would give me BSOD's *only* if I filled past a certain point, size-wise, even after a low-level format), faulty RAM, or a failing processor.  It *could* also be possible that a component somewhere on the motherboard has a bad connection, which would be beyond your ability to test/fix, and would be a bear to narrow down, regardless.  It may be time to take it into a reputable technician, and let them lose some hair over the matter.  They should have spare parts on hand to swap in & out to more quickly (a relative term) narrow down where the problem lies.  As it's intermittant, they may need to keep it for a while, so hopefully it's not your only machine.

[zuludelta]

Just throwing it out there, but you could be in need of a BIOS update. You might want to do a search for problems related to your particular motherboard model and see if other users have faced problems similar to yours. I was recently troubleshooting a PC (an HP model using a somewhat rare VIA mobo) that had symptoms similar to the ones you have... CPU usage running up to 99% at odd moments, fan kicking up a racket even when there's no apparent heavy processor load, etc. Turns out a simple BIOS update was the solution (well, not so simple actually: rare, HP-exclusive board that it was, it took more than a few hours of 'net legwork to find an "unofficial" community-created updated BIOS).

If you do decide to try a BIOS update, I recommend doing it with the help of somebody who's done it before and to back up your old BIOS and settings.