A user getting this message on a newly installed Windows 2008 on HP hardware was able to fix the problem by tweaking various hardware settings as instructed by HP engineers. These included updating the firmware for the controllers used by the servers and adjustments on CPU power management.
If the system is using an Intel processor, one may start the troubleshooting process by running the Intel Processor Diagnostic Tool (see EV100489
). However, several users reported that the tests were successful (no issues) even though the problem persisted.
From a newsgroup post, for an ASUS motherboard with an AMD CPU: "After three weeks of back and forth emails with MS tech support, here is the final answer that I got from them. Admittedly, my motherboard is slightly older but it worked great with XP and I've seen plenty of other people with much more recent AMD-based motherboards complaining about the same problem.
My hardare: ASUS a7n8x-e deluxe (nforce2 chipset), AMD Athlon xp 3200+, 1 GB Corsair (two dimms dual channel).
Email from MS tech support:
"I understand we have installed the latest BIOS update and AMD process update though they still too old. Since the current CPU is AMD Socket A and the Motherboard is ASUS Nvidia Nforce2 A7N8X-E, I checked the related information on their website. I am sorry to tell you that Nvidia claimed that nForce2 chipset has been discontinued as of date, since the product has reached the end of life stage. This is the reason they are unable to provide drivers with respect to Windows Vista operating system.
So these original BIOS and drivers may caused some conflict issue with Vista SP1. I also hope you could understand that there is little thing we can do on the system side to update these hardware BIOS and drivers. It is why you may received some warning prompt. I also submitted this information to our product group. They will check if there is available to release some fix tool to cover it, but as I said, since it is related to hardware, it is hardly to do something only on system side."
From a newsgroup post, by Microsoft support engineer:
"AMD provides a tool called MCAT (Machine Check Exception Analysis Tool). You can find it on the AMD site in the downloads/processor utilities section. Choose the version appropriate to your processor. It installs a command line utility. To use it, execute mcat on the command line. There are several ways to use it. Execute mcat /? to see the possible options. Hopefully it can shed some light on the source of your problems.
Modern processors such as yours support advanced self-diagnostic capabilities, such that they analyze their own behavior and are capable of recognizing conditions that "can't happen". When one of these not-possible conditions occur, the CPU raises a "machine check exception". Windows is simply retrieving the information provided by the processor and recording it in the event log.
You mention on this thread that you've tried many different processors. That would be my first recommendation. This eliminates the processor as a potential source of the problem. Several other possible causes still remain.
- Poor voltage regulation (i.e. power supply problem, voltage regulator malfunction, capacitor degradation)
- Damage due to power spikes
- Static damage to the motherboard
- Incorrect processor voltage setting in the BIOS (too low or too high)
- Permanent motherboard or power supply damage caused by prior overclocking
- Excessive temperature caused by insufficient airflow (possibly caused by fan failure or blockage of air inlet/outlet)
- Improper BIOS initialization (the BIOS configuring the motherboard or CPU incorrectly)
- Installation of a processor that is too much for your motherboard to handle (excessive power requirement, incompatibility)
- Defective hardware that may be drawing excessive power or otherwise disrupting proper voltage regulation
The TLB is a part of the processor that manages translation of virtual addresses to physical addresses (to implement advanced memory management required for a multitasking operating system). The processor uses "page tables" to map virtual addresses to physical addresses. The TLB caches information from the page tables so it won't have to keep looking up this information. The machine check exception in your post shows that, while the processor was attempting to fetch instructions to be executed, it detected invalid information in the TLB. This can lead to major software errors, so it raises a machine check exception. If windows doesn't crash, the processor may have detected a benign error, but still an error worth reporting since it "can't happen". It is highly unlikely that this sort of error could be caused by software (i.e. Windows), since the operation of the TLB is entirely automatic and managed by the processor itself.
Machine check exceptions are almost always caused by a hardware problem."
Error Type: TLB Error - As per ME961080
, this is caused by a bug in the Windows Hardware Error Architecture. A hotfix is available. See the article for details.