Well, not to disagree with you...but...
 
The motherboard in my system (ASUS A7N8X Deluxe) uses the nForce chipset.
 
Also, if I'm not mistaken - I've seen Intel users post similar problems in newsgroup threads about this topic. Hmm, I think in Google Groups somewhere, although I can't really remember specifics.
 
However, among people experiencing this issue - the vast majority do seem to be AMD users.
 
Thanks

Hmmm, how can you disagree with something that I said I THINK i may have 1 outta 2 correct. I really don't know.
 
I tend to just ignore the errors.

PROBLEM SOLVED!
 
Gentlemen...I solved this issue last night. After a few hours of testing on two different machines, I have perfect results from my "fix".
 
Here's how I fixed it. Keep in mind, this is on an AMD machine. I'm not sure how this affects Intel machines, or even if certain Intel machines have this issue to begin with as a result of MCA support.
 
1) Open Regedit
 
2) Browse to this key :
 
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\EnableMCA
 
3) Change the DWord value from a 1 (default) to a 0 (zero).
 
4) Reboot
 
5) Fixed
 
After doing this, I no longer have application popup 26 - Machine Check Regs on my AMD rigs (2 used for experimentation).
 
Good luck and if you're the paranoid type - back up your registry before changing this value. I happened to stumble across this key while entering my own custom registry tweaks on a newly installed OS for about the 20th time during the past two weeks
 
Happy Computing!

Interesting, I'll need to try this out.
 
Also, some further info found out Jsifaq, which relates to something very similar.
 
http://www.jsifaq.com/subd/tip1600/rh1648.htm
 
The Pentium and Pentium Pro processors provide a mechanism to detect and to report hardware-related problems such as memory parity errors and cache errors. To signal a hardware error, the processor signals the detection of a machine check error by generating a machine check exception (Interrupt 18). Windows NT simply reports the fact that the error occurred and displays parameters that you can use to decode the exception. Contact your hardware vendor or processor manufacturer for information regarding the Machine Check Architecture or consult the Intel Pentium Pro Family Developer's Manual - Volume 3: Operating System Writer's Manual.
 
Just some interesting reading.

Also, read the following:
 
Machine Check Architecture (MCA) is an internal architecture subsystem which detects and captures errors (i.e. faults) occurring within the microprocessor's logic.
 
This info was found through an intel site. So, I can only guess that both intel and amd have this feature, which the os reads. Perhaps that app popup just signifies that everything is good.
 
Does this happen on intel processors? I just now checked for this even on two machines here at work. Neither have this event in event viewer.
 
I've also checked this on an Athlon xp, sempron, amd 64 and amd 64(sempron). All of these computers show this popup.
 
Does MS correctly read the MC from intel cpus and not from amd? Or does windows mininterpret amd's version of the machine check.
 
Yes, this is also related to the linux world. Well, atleast now we know how to disable this "feature" in windows.

Originally posted by theefool:

Quote: Or does windows mininterpret amd's version of the machine check. 
 
I'm leaning that direction I think the smart money's riding on that as a very valid possibility.
 
Thanks for the info above

You can add me to the list of those with this problem.
 
Thanks to RBogan for the registry fix - I just went in and changed the value; has everyone seen an improvement on the number of lock-ups after changing the key?
 
I also went in today and did a fresh round of driver updates.
 
I do have the Via chipset that was discussed above...
 
I've gone round and round with this; its nice to find others that have shared in my misery.
 
Let me know how you have all made out.
 
Bo
 
********
Soyo kt400
amd athlon xp 2500
wintv pvr
geoforce fx 5200
win xp pro sp2
 
 
 

well, my machine froze again today - I'll check it out later - but I have a question - is it possible that whatever process/service (guessing its a process/service and not memory or hardware) that's causing this freeze does so before an event log can be generated?
 
thanks in advance for the edjumacation.
 
bo

I have this problem too..
 
Windows XP Pro SP2 - all updates
AMD AthlonXP 1600+ on Shuttle AK31 Rev. 3.1
VIA KT266A Chipset, 512 MB DDR-266
Creative SB Live PCI, Network Card, TV-Card,
DVD-ROM, DVD-RW, 2xHDD, FDD, 1xUSB used,
Gainward GF5900 AGP
VIA 4in1-Driver 4.56
 
First I thought it was a problem about my ram, but
memtest didn't find anything. My next choice was
the power supply - a 300W one.. a new one is ordered
and iam waiting for it. Because of I have to wait for
my new ps I looked a bit around and the next thing I
suspected to cause the problems was the NoExecute Feature
of the built-in Windows SP2 Firewall. I switched it off
in the boot.ini - because the CPU doesn't have this
feature. No changes from this, but a good feeling to
know it's disabled
 
Problem: Machine Check: Regs (sometimes without Regs)
on every cold boot. Some crashes in unspecified situations
like listen mp3s and surfing the internet, using Trillian (ICQ,MSN,Yahoo) and so on..
 
Since I have changed the case to a Chieftec Big-Tower with
proper cooling abilities, I haven't had any crashes - hope this is going on this way. BTW, I don't think it's the best way to disable the MCA feature, but that's an everyone´s selfmade decision.
 
Results:
bad RAM - no
bad PSU - maybe for crashes, but not for Machine Check "Information"
bad CPU - no, everything was fine until SP2
bad VIA Driver - maybe, has anyone checked it with older drivers ? (to avoid the crashes and/or the Machine Check thing)
 
If I find any other issue to solve the problem, I will post it here.
 
@bo_bo: If you find no Event for this (i never did) it's an error that occurs in the hardware or in a high kernel level. If it is anything else, Windows will give out a BSOD/reboot, and you have the Event for a saved Minidump to check. AFAIK..
 
Best regards
mainhack

I wonder whether you have optical usb mouse as I have logitech MX300.
[Edited by steadyeye on 2005-10-09 13:05:15]
 

Sorry, the mouse issue must not be related to this problem.
I removed optical mouse, its driver but my computer still freezes up with machin check log msg.

I solved the problem in my case.
 
I had both computer freezing and 'machine check' log msg.
 
I tried to find the cause by any kinds of methods.
 
Finaly, I pluged off the power from M/B. Then i suprised that power 2 plastic pins connecting to M/B had been burend among total 20pins.
 
I guess irregular power was the cause. and I have experienced no freezing and no log msg after replacing burned power to another power.
 
If freezing and log msg occur, i will reply here again.
 
Good Luck to all

I have what seems to be the same issue.
 
Unfortunately, I believe there is a problem with the machine, and this message is not just for "Information". The problem is that the computer runs fine for a couple of weeks, then when I boot the computer, Windows will not load and display "The file c:\windows\system32 is missing or invalid". I need to restore the registry and then it works again for a couple of weeks.
 
I'm running a AMD Duron 1200 on a VIA-KT133 chipset (I'm not 100% sure about the KT-133, but it's definitely a VIA). Unfortunately, the machine is at home, and I'm at work. On every boot, I'm getting 4 instances of "Application Popup" Machine Check logs in my event viewer. Only 1 of the 4 has "Machine Check: Regs", the other 3 are just "Machine Check". This is on a clean install of WindowsXP with Service Pack 1. No changes to the system have been made since the OS install (no other software has been installed either).
 
These are the only "problems" that show up in the event viewer. Every other event is normal. The strange thing is that if I run a chkdsk /f, the very next boot doesn't show the Machine Check in the event viewer. It then shows up on the following boot and every subsequent boot.
 
I've disconnected all hardware (pci cards, drives, etc..) except for the HD, ram, and processor. I've switched out the video card and still get the same errors. I've ran memtest on the ram and it is okay. Tonight, I'll be running diagnostics on the HD to verify the HD is 100% okay.
 
The point of this post is to say that I think, although it might have not manifested itself, that the "Machine Check: Regs" information in the event viewer may be because of a critical error somewhere. I can't find much helpful information anywhere else. Also, I'm going to try to disable write behind caching on the HD. I've read that the Machine Check message may display when the system powers off before windows is done writing to disk on shut down.
 
Anyway, if anyone has an idea why I get the "The file c:\windows\system32 is missing or invalid" after a couple of weeks on my machine, I'd appreciate any ideas. I figured that if I fix the "Machine Check: Regs" problem, that I will not get the system32 missing or invalid error.

I've had lots of experience with this error message on many computers, and I don't think it can really be blamed on AMD or VIA. I took apart and rebuilt about 40 older pentium systems and had this problem pop up quite often. Power issues could definitely be causing it, like someone else just said... (the power supplies on these machines were quite prone to sudden death.
However, to me it seems more like a memory issue. Not necessarily bad memory sticks, but possibly wrong settings. I've gotten the errors to go away (some of the time anyway) by changing the memory settings (timing, etc) and moving the sticks around. Messing with other settings in BIOS has worked for me as well.
One thing I have noticed is that when I have a computer thats getting those "machine check: regs" errors, windows seems to slowly but gradually get corrupted and almost always needs to be reformatted, although it can take months to a year before it gets that bad. This is what leads me to believe it could be a memory related error message... even if your memory sticks pass memtest etc, maybe they just don't work well with your particular motherboard?
 
Just my thoughts,
Peace

I've been running my AMD Duron 1100 / Gigabyte GA-7VAX (VIA KT-400 chipset) for a couple years now, through several peripheral hardware upgrades (disk, cdrom, cpu and case fans, etc.) and have always had the same problem. I just upgraded to WinXP so I can now see the event log messages ("Machine Check" and "Machine Check:Regs" repeated 4 times) but the appearance of the problem for me was the same under Win98SE.
 
It only ever seemed to happen under absolute "cold" bootup. This means leaving the PC off for a long time (like overnight). Never happened on a reset/reboot or even complete power off and power on after a few minutes. Also, if I cold booted and went into the bios setup for a minute to let the system "warm up" the problem didn't happen.
 
Under both Win98 and also XP, the most regular appearance was McCaffee Virus scan startup reporting "corrupt signature files" or something similar. XP sometimes reports some other errors too (i.e., corrupt registry recovered). The problem always goes away after rebooting. A 2nd reboot is sometimes, but usually not needed.
 
I always suspected the power supply but I upgraded that and the problem still happened. I also suspected a mechanical (thermal) contact issue but I've seated/unseated/reseated cards, cpu, cables many times through many upgrades and still the same problem. Once the system warms up and boots cleanly, never a problem. No memory test or cpu test ever indicated a problem.
 
Actually, now that I have XP, my system boots up much more quickly and the dreaded scandisk that used to take forever seems to be much rarer so it's become much less of an annoyance for me.
 
Still, like everyone else here, I'd like to know what's actually causing these machine checks and whether it's really ok to simply ignore them (suppress in registry)? And does anyone know what type of processing McCaffee viruscan might do at initialization that causes it to see this problem when everything else going on at boot up seems to be ok?
 
alfei

Just a quick update to my previous post.
 
I did some more experimenting and the McCaffee startup problem may not be directly related to the "machine checks". I only get McCaffee errors on an absolute cold bootup (as described). The machine checks appear in the log on *every* cold bootup, even if the power was off for only 10 seconds and no McCaffee errors show up.
 
alfei

Just a question - errors like this happen especially with a machine that was overclocked - not just because it was overclocked but because of the excessive heat. The thing is, the heat is not distributed over the CPU evenly meaning that certain areas of it are exceedingly hot and others are "normal." So, here's the question, have you swapped out your CPU with a new one and had the problem disappear? Could it be that the alloy within the CPU is changing the electrical flow or causing it become erratic either changing the timing or "misloading" the internal cache bacause after being sufficiently "cold" the alloy has to be rewarmed to operate within spec again? Could it even be one of the pins has oxidized? The question is - have you changed your CPU to see if it is the problem?

Thanks for the suggestions.
 
I've never run with a different CPU (been wanting to upgrade to an AthlonXP-Barton but the prices keep going UP) but about a year ago I upgraded my CPU heatsink/fan. I removed the CPU and cleaned the surface (thermal compound was dried out) and applied new thermal compound in the process. The problem, though, has always happened as far as I can remember. Also, I've never overclocked this system.
 
That upgrade was for heat concerns (after I started using MBM) and included adding a case fan or two. Since then, I've been more careful about keeping power & other cables neat to improve air circulation inside the case. I have a desktop case which was chosen based on physical space requirements in my apartment before I paid any attention to heat build up.
 
It seems the more monitors I have (MBM, now WinXp logs) the more there is to worry about .

I joined just to let you know that this is not AMD/Via specific. I have gotten the same messages as you after a cold system crash on my rig.
I have Pentium 930, 2x512MB DDR2-667 Geil dualchannel kit, eVGA 7800GT, Asus P5LD2 mainboard , WD2500JS 250GB HDD, and 550W PSU.
It happens to me while watching DivX video (only if I use 6.x codec) and while playing Battlefront. It didn't happen on the same system with Pentium 630, all the rest being the same. Nothing is overclocked and nothing is crossing 60°C.
 
MCA (machine check architecture) is designed to detect and report machine (hardware) errors such as bus errors, ECC errors, parity errors, cache errors, and TLB errors. Machine check exception (also called fault) is being generated when such an error occurs and the CPU must be reset in order to continue. When this exception happens, OS suspends any software running and records exception information from CPU MSR registers and then it crashes/reboots.
 
It beats me what could be wrong with my computer though because I have used all branded components to put it together.