Tool AT-W23REV. 2026-A

Blue screen help: what does the stop code mean, and what helps?

A blue screen shows a stop code at the bottom, such as CRITICAL_PROCESS_DIED or 0x000000EF. It tells you where Windows stopped, not always why. According to Microsoft, third-party drivers are behind around 70 per cent of cases and hardware behind 10 per cent. The blue screen help recognises all 379 codes in the Microsoft reference, even with typos, and ranks causes and steps according to your answers. It also tells you when it is better to stop.

  • free
  • no sign-up
  • runs in the browser

Enter the stop code, read off causes and steps

Tool · Look up a stop code

The stop code names the place where Windows stopped, not always the cause. The ranking below is a reasoned narrowing-down; only an analysis of the crash logs gives certainty.

Step 1Which stop code is on the screen?

Name or hex number, with or without 0x and leading zeros; typos are not a problem. You may paste in text from a photo or from Event Viewer in full. It also works without a code.

The module is the file Windows names under the code, usually ending in .sys. Optional.

Step 2What changed before the first blue screen?
Step 3Since when has it been happening?
Step 4How often?

Code recognised

—

All codes in the Microsoft reference

Likely causes

    Steps in order

    Ordered by likelihood and effort. After each step, work normally for a while; if no blue screen comes back, you are done.

      Stop here and get help

        Codes according to the Microsoft reference, as of 14 November 2026. The crash can only be narrowed down further with the memory dumps. The easiest way to send us the copied result is the contact form.

        It runs in your browser. Nothing is sent or stored.

        How to read the blue screen

        Windows stops when it detects a situation in which it can no longer safely carry on, for example because data would otherwise be damaged. Microsoft calls this a bug check. Three things on the screen then matter.

        1. The stop code. At the bottom of the screen, as a name in capitals with underscores. This is the value you enter above.
        2. The module, if shown. Below that, Windows sometimes names a file, usually ending in .sys. It was running when Windows stopped and is the best clue to a driver.
        3. The progress percentage. Windows is writing the memory dump from which the cause can later be read. Let it finish.

        From Windows 11 version 24H2 the message can be black instead of blue; the code is shown further down. If the screen disappeared too quickly, you will find the crash in Event Viewer under Windows Logs > System, with the code as a hex number. The memory dumps are in the folder C:\Windows\Minidump.

        3CRITICAL_PROCESS_DIED1beispiel.sys21 · STOP CODE · ENTER THIS AT THE TOP2 · MODULE · DRIVER FILE, IF SHOWN3 · PERCENT · WINDOWS IS WRITING THE DUMPSCHEMATIC · LAYOUT VARIES BY WINDOWS VERSION

        Where blue screens come from

        Microsoft has analysed what is behind the crashes that reach it. The result is clear: most of them trace back to drivers that do not come from Microsoft. A driver runs in the core of Windows, and an error there cannot simply be caught the way it can in a program.

        Research has confirmed this for years. Under Windows XP, 85 per cent of reported failures were caused by drivers Swift et al. 2005. In operating system kernels, drivers contained three to seven times as many errors as the rest of the code, and fresh code was more error-prone than older code Chou et al. 2001. Ten years later, drivers still contained the most errors, simply because they make up most of the code Palix et al. 2011. That is why, with no other information, the tool puts a driver first, and a freshly updated driver counts as the prime suspect.

        Causes of stop errors according to Microsoft
        ShareCause
        70 %Third-party driver code
        10 %Hardware problems
        5 %Microsoft code
        15 %unknown, because the memory was too badly damaged for analysis

        Source: Microsoft Learn, “Advanced troubleshooting for Stop error or blue screen error issue”, as of 12 February 2026.

        Programs with their own driver, and updates

        Not every driver belongs to a device. Third-party antivirus, VPN, backup and tuning programs, or the anti-cheat software of some games, bring their own drivers. Under Windows XP, crashes caused by drivers fell from 2004 to 2006 in all classes except two: antivirus and storage. Antivirus drivers are constantly being reworked, and that keeps producing waves of new errors Glerum et al. 2009. The same analysis also shows why updates help: one manufacturer distributed a fix for a processor error via Windows Update, and within two days the crash reports fell to a fifth.

        Updates can, however, also be the trigger themselves. In a survey of 50 administrators, two thirds estimated that 5 to 10 per cent of the updates they installed cause problems Crameri et al. 2007. If you say the blue screen began after an update, you therefore get as the first step after the restart: uninstall the update. For every code, Microsoft advises checking the last change first; a change you name therefore outranks every other assessment in the tool.

        When hardware is the cause

        Hardware accounts for ten per cent, and such errors are rarely one-offs. In the crash data from around 950,000 consumer PCs, the chance of a crash caused by the processor was 1 in 190. After the first such crash it was 1 in 3.3. Overclocked processors crashed 4 to 20 times as often, laptops less often than desktops Nightingale et al. 2011. Memory is similar: over 8 per cent of the memory modules in Google’s data centres had errors every year, and permanent faults predominated, not one-off glitches Schroeder et al. 2009 Hwang et al. 2012 Sridharan and Liberty 2012. In Facebook’s servers, most memory errors came not from the chips but from the memory controller and channel Meza et al. 2015; that is why the tool advises testing modules in another slot as well.

        For the assessment this means: if you say “for weeks” or “daily”, memory, drive and heat move further up. A faulty memory module often produces quite different codes, incidentally, because it corrupts data in changing places. If you keep getting different codes, that is itself a clue.

        The 61 most common stop codes in plain language

        The tool describes these codes individually, based on the relevant Microsoft page. It also recognises the other codes among the 379 and assigns them to a group by their name. Each name links to the Microsoft page; it is written for developers, but it gives the parameters in case someone analyses the memory dumps.

        Common stop codes by number
        CodeNameMeaningGroup
        0x00000001APC_INDEX_MISMATCHA driver or file system did not cleanly finish internal processes in the Windows kernel.Drivers
        0x0000000AIRQL_NOT_LESS_OR_EQUALWindows or a driver accessed an invalid memory address at a moment when that is not allowed.Drivers
        0x00000018REFERENCE_BY_POINTERA driver released a system object more often than it requested it.Drivers
        0x00000019BAD_POOL_HEADERA memory area of the Windows kernel (pool) is damaged.Memory management
        0x0000001AMEMORY_MANAGEMENTA severe error occurred in the memory management of Windows.Memory management
        0x0000001EKMODE_EXCEPTION_NOT_HANDLEDA program in the Windows kernel, usually a driver, triggered an error that nobody caught.Drivers
        0x00000024NTFS_FILE_SYSTEMError in ntfs.sys, the driver through which Windows reads and writes NTFS drives. Back up your data first.Hard drive and file system
        0x0000003BSYSTEM_SERVICE_EXCEPTIONAn error occurred when a program passed into the Windows kernel, usually an invalid memory access.Windows kernel
        0x00000044MULTIPLE_IRP_COMPLETE_REQUESTSA driver tried to complete a request that had already been completed.Drivers
        0x0000004EPFN_LIST_CORRUPTThe list used to manage RAM is damaged.Memory management
        0x00000050PAGE_FAULT_IN_NONPAGED_AREAWindows accessed invalid system memory, often memory that had already been freed.Memory management
        0x0000005ACRITICAL_SERVICE_FAILEDA service that Windows needs in order to start has failed.Windows startup
        0x0000006BPROCESS1_INITIALIZATION_FAILEDWindows could not initialise at startup. Back up your data first.Windows startup
        0x00000074BAD_SYSTEM_CONFIG_INFOThe registry is faulty: the system part is damaged, or important entries are missing.Windows startup
        0x00000077KERNEL_STACK_INPAGE_ERRORA page of kernel data could not be read from the page file into memory. Back up your data first.Hard drive and file system
        0x0000007AKERNEL_DATA_INPAGE_ERRORKernel data could not be read from the drive into memory. Back up your data first.Hard drive and file system
        0x0000007BINACCESSIBLE_BOOT_DEVICEAt startup, Windows lost access to the system partition. Back up your data first.Windows startup
        0x0000007ESYSTEM_THREAD_EXCEPTION_NOT_HANDLEDA system process in the Windows kernel triggered an error that was not caught.Drivers
        0x0000007FUNEXPECTED_KERNEL_MODE_TRAPThe processor reported an exception that the kernel could not catch.Hardware
        0x00000080NMI_HARDWARE_FAILUREThe hardware has reported a severe error.Hardware
        0x0000008EKERNEL_MODE_EXCEPTION_NOT_HANDLEDA program in kernel mode triggered an error that was not caught. Microsoft lists this code as very common.Drivers
        0x0000009CMACHINE_CHECK_EXCEPTIONThe processor has reported a severe hardware error (machine check).Hardware
        0x0000009FDRIVER_POWER_STATE_FAILUREA driver did not respond in time, or responded incorrectly, when switching to a power-saving state.Power management
        0x000000A0INTERNAL_POWER_ERRORThe power management of Windows has detected a severe error.Power management
        0x000000A5ACPI_BIOS_ERRORThe BIOS/UEFI of the computer does not fully comply with the ACPI standard for power and device management.Power management
        0x000000BEATTEMPTED_WRITE_TO_READONLY_MEMORYA driver tried to write to write-protected memory.Drivers
        0x000000C2BAD_POOL_CALLERA driver requested kernel memory incorrectly or released it twice.Memory management
        0x000000C4DRIVER_VERIFIER_DETECTED_VIOLATIONDriver Verifier has detected a driver error.Drivers
        0x000000C5DRIVER_CORRUPTED_EXPOOLA driver has damaged the system memory pool.Memory management
        0x000000CAPNP_DETECTED_FATAL_ERRORPlug and Play management has detected a severe error, probably caused by a device driver.Devices and USB
        0x000000CEDRIVER_UNLOADED_WITHOUT_CANCELLING_PENDING_OPERATIONSA driver was unloaded without ending its pending operations.Drivers
        0x000000D1DRIVER_IRQL_NOT_LESS_OR_EQUALA driver accessed pageable or invalid memory at the wrong time.Drivers
        0x000000D5DRIVER_PAGE_FAULT_IN_FREED_SPECIAL_POOLDriver Verifier caught a driver accessing memory that had already been freed.Drivers
        0x000000E2MANUALLY_INITIATED_CRASHThe crash was triggered deliberately, using a debugger or a key combination.triggered deliberately
        0x000000EATHREAD_STUCK_IN_DEVICE_DRIVERA device driver is stuck in an endless loop, waiting for the hardware.Image
        0x000000EDUNMOUNTABLE_BOOT_VOLUMEWindows could not mount the boot drive at startup. Back up your data first.Windows startup
        0x000000EFCRITICAL_PROCESS_DIEDA process that Windows cannot do without was ended, such as csrss.exe, wininit.exe or winlogon.exe.Windows kernel
        0x000000F4CRITICAL_OBJECT_TERMINATIONA process or thread needed for operation ended unexpectedly.Windows kernel
        0x000000F5FLTMGR_FILE_SYSTEMSevere error in the filter manager, through which programs hook into every file access.Hard drive and file system
        0x000000F7DRIVER_OVERRAN_STACK_BUFFERA driver overwrote a buffer area; Windows stops before damage results.Drivers
        0x000000FCATTEMPTED_EXECUTE_OF_NOEXECUTE_MEMORYCode was supposed to run in a memory area that is locked for that purpose.Drivers
        0x000000FEBUGCODE_USB_DRIVERError in a USB driver.Devices and USB
        0x00000101CLOCK_WATCHDOG_TIMEOUTA processor core stopped responding.Hardware
        0x00000109CRITICAL_STRUCTURE_CORRUPTIONWindows has detected modified kernel code or kernel data.Windows kernel
        0x0000010DWDF_VIOLATIONA driver built on the Windows driver framework made an error.Drivers
        0x0000010EVIDEO_MEMORY_MANAGEMENT_INTERNALThe graphics memory management cannot recover from an error.Image
        0x00000113VIDEO_DXGKRNL_FATAL_ERRORThe DirectX graphics kernel of Windows has detected a violation.Image
        0x00000116VIDEO_TDR_FAILUREThe graphics card stopped responding, and restarting the graphics driver failed.Image
        0x00000119VIDEO_SCHEDULER_INTERNAL_ERRORThe graphics scheduler of Windows has detected a severe error.Image
        0x00000124WHEA_UNCORRECTABLE_ERRORThe hardware has reported a severe error that could not be corrected.Hardware
        0x0000012BFAULTY_HARDWARE_CORRUPTED_PAGEWindows has detected damaged memory that only direct hardware access can have caused.Memory management
        0x00000133DPC_WATCHDOG_VIOLATIONA driver took too long and blocked the system as a result.Drivers
        0x00000139KERNEL_SECURITY_CHECK_FAILUREThe Windows kernel has detected a damaged internal data structure.Windows kernel
        0x0000013AKERNEL_MODE_HEAP_CORRUPTIONThe kernel’s memory management has detected damaged memory blocks.Memory management
        0x00000144BUGCODE_USB3_DRIVERError in a USB 3 driver.Devices and USB
        0x00000154UNEXPECTED_STORE_EXCEPTIONA component of the Windows memory management caught an unexpected error.Memory management
        0x00000164WIN32K_CRITICAL_FAILUREWin32k, a core part of Windows for the user interface, has encountered a critical error.Windows kernel
        0x00000192KERNEL_AUTO_BOOST_LOCK_ACQUISITION_WITH_RAISED_IRQLA driver requested a lock at a time when that is not allowed.Drivers
        0x000001C8MANUALLY_INITIATED_POWER_BUTTON_HOLDThe power button was held down for a long time; Windows wrote a memory dump first.triggered deliberately
        0x000001D5DRIVER_PNP_WATCHDOGA driver did not complete a Plug and Play operation within the time allowed.Devices and USB
        0xC000021AWINLOGON_FATAL_ERRORWinlogon or CSRSS has crashed; Windows cannot run without these processes.Windows startup

        Numbers and names from Microsoft’s Bug Check Code Reference, retrieved on 14 November 2026. According to Microsoft, 0x1000007E (SYSTEM_THREAD_EXCEPTION_NOT_HANDLED_M), 0x1000007F (UNEXPECTED_KERNEL_MODE_TRAP_M), 0x1000008E (KERNEL_MODE_EXCEPTION_NOT_HANDLED_M) and 0x100000EA (THREAD_STUCK_IN_DEVICE_DRIVER_M) have the same meaning as 0x0000007E, 0x0000007F, 0x0000008E and 0x000000EA.

        The steps at a glance, starting with the smallest intervention

        The order follows Microsoft’s advice to start with the least disruptive option. The tool moves forward what fits your code and your answers and leaves out what does not; for a hardware fault, for example, resetting Windows. The restart still always comes first. It fixes many faults, but not all: on Windows servers in an early field study, restarts did not always solve the problem, and failures occurred in clusters Kalyanakrishnam et al. 1999.

        Steps, effort and what happens to your files
        Stephelps withEffortYour files
        Restart and note the codeone-off glitches5 minuteskept
        Install Windows updatesknown driver and processor errors15 to 30 minuteskept
        Uninstall the latest updateblue screen since an update10 to 20 minuteskept (Microsoft)
        Roll back or renew the driverdrivers, including the one named on the screen15 to 30 minuteskept
        Remove recently installed programs with a system driverantivirus, VPN, tuning, anti-cheat15 to 30 minuteskept
        Unplug new hardware againnew parts and devices10 minuteskept
        Undo overclocking and memory profileoverclocking, memory profile, BIOS change10 minuteskept
        Test the memorymemory and slots30 to 60 minuteskept
        Check and repair system filesdamaged Windows files20 to 40 minuteskept
        Check the hard drive or SSD, after backing upread errors, cables, drive20 minutes to 2 hoursback up first
        Test in Safe ModeNarrowing down: driver or Windows?15 to 30 minuteskept
        System Restore to a point before the problemchanges since the restore point30 to 60 minuteskept, programs installed afterwards are gone
        Reset Windows, as the last software stepstubborn software faults1 to 3 hourskept with “Keep my files”, programs are gone

        Effort is a rule of thumb. Menu paths follow Windows 11, as of 14 November 2026; under Windows 10 some items have different names. Whether your Windows 10 still receives updates is shown by the Windows 10 check.

        Two traps when restoring

        • Point-in-time restore. From Windows 11 24H2, Microsoft lists it first for a PC that no longer starts. According to Microsoft, it rolls back the whole PC, including personal files. If you have saved anything since the restore point, you are better off using the classic System Restore or backing up beforehand.
        • BitLocker. For most options in the recovery environment, Windows asks for the 48-digit recovery key if the drive is encrypted. It is often held in the Microsoft account. Without it, the data is lost after a reset, and Microsoft cannot recover it.

        When to stop and get help

        Most of the steps above are harmless. With a drive that already has read errors, that is different. After a first crash caused by the disk system, around one in three or four PCs crashed again, most within ten days of running time; the researchers advise replacing the drive straight away rather than waiting Nightingale et al. 2011. Even one unreadable sector can cost data without a backup Schroeder et al. 2010, and a rising number of reallocated sectors signals a failure Ma et al. 2015. The hard drive check shows how your drive is doing.

        • The drive clicks, grinds or is not recognised. Switch off. Every further start reads from the failing drive, and chkdsk /r even more so. If the data exists nowhere else, this is a case for data recovery; the guide Hard drive failed gives first pointers.
        • BitLocker asks for the key, and it cannot be found. Do not reset and do not reinstall.
        • The memory diagnostic reports errors, or the crash persists despite clean fans and BIOS default settings. Then a component is faulty; software steps no longer help. The Repair or replace? tool calculates whether the repair is worthwhile.
        • There is a smell of burning, the device gets unusually hot, the battery is swelling. Switch off immediately and disconnect from the power. Why laptops get hot is explained in the guide Laptop loud and hot.
        • The blue screen comes back after all the steps. Then the memory dumps need analysing. Even Microsoft does not group crash reports by code alone, but compares the call stacks in the dump Dang et al. 2012; only they show which driver was involved.

        If the computer no longer even gets as far as the blue screen, the guide PC won’t start can help. If it runs but is sluggish and keeps stuttering, take a look at PC slow: what to do. We analyse the memory dumps by remote support, without you having to bring the computer in; if a part has to be replaced, see Getting a computer repaired.

        What the research says about crashes

        The order in the tool is based on Microsoft’s information and on 14 vetted papers. Six findings support it; the full list with verification notes is further down.

        • Drivers are the most common cause. 85 per cent of failures under Windows XP Swift et al. 2005, three to seven times as many errors as in the rest of the kernel Chou et al. 2001.
        • New code is more error-prone. The newest files had up to twice as many errors as the oldest Chou et al. 2001; after a driver update, the driver therefore comes first.
        • Hardware faults come back. After the first crash, a second one from the same cause is up to a hundred times more likely Nightingale et al. 2011. Frequency is therefore a question of its own in the tool.
        • Memory errors are mostly permanent. That is why the memory test is worthwhile for recurring crashes Schroeder et al. 2009 Hwang et al. 2012.
        • Updates fix problems and cause them. Fixes arrive via Windows Update Glerum et al. 2009, and some updates cause trouble themselves Crameri et al. 2007.
        • The code is not a diagnosis. Even Microsoft does not classify crashes by code alone, but compares the call stacks in the dump Dang et al. 2012.

        Using this tool from an AI assistant

        Assistants such as Claude or ChatGPT that speak the Model Context Protocol (MCP) can call this page’s calculation directly instead of rebuilding it. The server is public, needs no sign-in and stores nothing. The answer always includes the link to this page, the date of the data and the limits of the result.

        Endpoint
        https://aureum-tech.com/mcp
        Tool
        werkzeug_bluescreen_hilfe
        Transport
        Streamable HTTP, read-only
        Inputs
        NameRequiredMeaning
        stoppcodenoStop code as a name (CRITICAL_PROCESS_DIED, also with spaces or typos) or as a hex number (0x000000EF, EF, 0xef). Pasted text from the blue screen or from Event Viewer (“The bugcheck was: 0x0000009f …”) also works. Leave empty if the code is unknown.
        modulnoOptional: file that appeared on the blue screen below the code, e.g. nvlddmkm.sys or ntoskrnl.exe.
        ausloesernoWhat changed before the first blue screen: nichts = Nothing, as far as I know; update = A Windows update; treiber = A new or updated driver; hardware = New hardware or a new device; programm = A new program; bios = BIOS setting or overclocking. If omitted: nichts.
        seitnoSince when: heute = For the first time today; tage = For a few days; wochen = For weeks or longer. If omitted: tage.
        haeufigkeitnoHow often: einmal = Once; mehrmals = Several times; taeglich = Daily or more often; startet_nicht = Windows no longer starts. If omitted: mehrmals.

        Questions where an assistant should call it

        • “My PC shows a blue screen with CRITICAL_PROCESS_DIED, what should I do?”
        • “What does the stop code 0x00000124 mean, and is my hardware faulty?”
        • “Since the last Windows update I get DRIVER_IRQL_NOT_LESS_OR_EQUAL with nvlddmkm.sys, what helps?”
        • “Windows no longer starts, blue screen INACCESSIBLE_BOOT_DEVICE. Is my data at risk?”

        How the connection works and what MCP can do in your own business is explained on the page about MCP integrations.

        Common questions about blue screens and stop codes

        What does the stop code on the blue screen mean?

        The stop code tells you where Windows stopped so that nothing worse could happen. It appears at the bottom of the screen as a name such as CRITICAL_PROCESS_DIED, and in Event Viewer as a hex number such as 0x000000EF. Microsoft lists 379 such codes. The code narrows things down; it does not always name the cause. According to Microsoft, third-party drivers are behind 70 per cent of cases and hardware behind 10 per cent.

        Is a blue screen serious if it only happened once?

        Usually not. According to Microsoft, a restart fixes the problem in most cases. Note down the code and the date, and install the updates. If the same code comes back, it is worth looking into: after a crash caused by hardware, a second one is far more likely than the first was. The exception is codes that point to the hard drive, such as KERNEL_DATA_INPAGE_ERROR. In that case, back up your data after the very first time.

        Where do I find the stop code if the screen disappeared too quickly?

        In Event Viewer under Windows Logs > System, at the entry for the time of the crash. The code appears there as a hex number, and you can enter that here. The memory dumps are in the folder C:\Windows\Minidump. If the computer keeps restarting, you can choose “Disable automatic restart on system failure” in the startup settings; the blue screen then stays on screen.

        What does CRITICAL_PROCESS_DIED mean?

        A process that Windows cannot do without was ended, such as csrss.exe, wininit.exe or winlogon.exe. As a number, the code is 0x000000EF. With no further information, the tool gives as the most likely causes “Damaged system files or registry”, “A device driver” and “A Windows update”. If it began after an update, “A Windows update” comes first, and the first step after the restart is to uninstall it.

        Is the hardware faulty if I get WHEA_UNCORRECTABLE_ERROR?

        Not necessarily, but the hardware has reported an error it could not correct. Microsoft names heat, overclocking, memory and a failing processor; a driver is behind it less often. If the code has been appearing daily for weeks, the tool puts “Heat or power supply” and “Memory (RAM) or slot” first. Check the fans and overclocking first, then the memory. If it carries on, a component has to be replaced.

        Can a blue screen destroy my data?

        Anything not yet saved at the time of the crash is lost. It becomes dangerous when the code points to the drive: in an analysis of around 950,000 PCs, roughly one in three or four with disk-system errors crashed again after the first time, mostly within ten days of uptime. In that case, back up first rather than spending hours on repairs. If Windows asks for the BitLocker recovery key, do not reset the PC without that key.

        Why is my screen black instead of blue?

        From Windows 11 version 24H2 the message can be black; Microsoft speaks of blue screen and black screen errors, and the stop code is still shown lower down. A black screen asking you to release the power button belongs to MANUALLY_INITIATED_POWER_BUTTON_HOLD. A black screen with no code at all is something else, for which Microsoft has its own guide.

        Where does the blue screen help stop?

        It recognises every code in the Microsoft reference and ranks causes and steps according to Microsoft’s information, studies and your answers. Which driver or component it really was is only shown by the memory dump. We analyse that, remotely or on site in Rodewisch and the Vogtland, and fix the cause.

        What the tool is based on

        The guide values and notes above are based on these 14 scientific papers. Each one has been checked via its DOI or the publisher’s page (most recently on 14 November 2026): it exists, it has not been retracted, and its findings still hold for this question today.

        1. Swift, M. M.; Bershad, B. N.; Levy, H. M. (2005): Improving the reliability of commodity operating systems. ACM Transactions on Computer Systems 23 (1), pp. 77–110

          Finding Device drivers remain a main cause of system failures: under Windows XP, 85 per cent of the failures reported at the time were caused by drivers. The paper proposes isolating drivers in the kernel so that their errors do not take the whole system down with them.

          In the tool With no other information, a driver comes first among the causes in the tool.

        2. Chou, A.; Yang, J.; Chelf, B. et al. (2001): An empirical study of operating systems errors. Proceedings of the 18th ACM Symposium on Operating Systems Principles (SOSP), pp. 73–88

          Finding An automatic code check of the Linux and OpenBSD kernels found three to seven times as many errors in device drivers as in the rest of the kernel. The newest files had up to twice as many errors as the oldest; code becomes more stable over time.

          In the tool A freshly installed or updated driver moves forward in the tool, and the step “Roll back or renew the driver” comes early.

        3. Palix, N.; Thomas, G.; Saha, S. et al. (2011): Faults in Linux: ten years later. Proceedings of the 16th International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS), pp. 305–318

          Finding Ten years after Chou et al., drivers still make up a large part of the Linux kernel and contain the most errors. Their error density is now lower than that of other parts such as file systems, however.

          In the tool The page explains that drivers are so often to blame mainly because there are so many of them, and not every driver is bad.

        4. Glerum, K.; Kinshumann, K.; Greenberg, S. et al. (2009): Debugging in the (very) large: ten years of implementation and experience. Proceedings of the 22nd ACM Symposium on Operating Systems Principles (SOSP), pp. 103–116

          Finding Windows error reports sort billions of crashes into groups and steer fixes back to the devices. Under Windows XP, crashes caused by drivers fell from 2004 to 2006 in all classes except antivirus and storage; antivirus drivers are constantly being reworked. A processor fix via Windows Update cut crash reports to a fifth within two days, and a driver manufacturer halved its share of all kernel crashes in five months.

          In the tool Installing updates comes early in the sequence of steps, programs with their own driver such as antivirus are a cause of their own, and the stop code counts only as a group, not as a diagnosis.

        5. Dang, Y.; Wu, R.; Zhang, H. et al. (2012): ReBucket: A method for clustering duplicate crash reports based on call stack similarity. Proceedings of the 34th International Conference on Software Engineering (ICSE), pp. 1084–1093

          Finding To group crash reports of the same error, the method compares the call stacks from the reports; for five Microsoft products it assigns them more accurately than the previous rules.

          In the tool Justifies the limit of the tool: only the call stack in the memory dump shows the culprit, not the stop code alone.

        6. Nightingale, E. B.; Douceur, J. R.; Orgovan, V. (2011): Cycles, cells and platters: an empirical analysis of hardware failures on a million consumer PCs. Proceedings of the 6th European Conference on Computer Systems (EuroSys), pp. 343–356

          Finding Crash data from around 950,000 consumer PCs: hardware faults come back. After a crash caused by the processor, the chance of the next was 1 in 3.3 instead of 1 in 190; with errors in the disk system, 86 per cent of repeats occurred within ten days of running time, and the authors advise replacing the drive straight away. Overclocked processors crashed 4 to 20 times as often, laptops less often than desktops.

          In the tool Repeated crashes and “for weeks” move hardware causes higher up; for drive codes the tool warns at the very first time, and undoing overclocking is a step of its own.

        7. Schroeder, B.; Pinheiro, E.; Weber, W.-D. (2009): DRAM errors in the wild: a large-scale field study. Proceedings of the 11th International Joint Conference on Measurement and Modeling of Computer Systems (SIGMETRICS), pp. 193–204

          Finding Two and a half years of measurements in Google’s server fleet: memory errors are orders of magnitude more common than assumed, with over 8 per cent of memory modules affected each year. Permanent hardware faults predominate; temperature plays a surprisingly small role in the field.

          In the tool For recurring crashes with changing codes, memory moves forward, and the memory test is a fixed step.

        8. Hwang, A. A.; Stefanovici, I. A.; Schroeder, B. (2012): Cosmic rays don't strike twice: understanding the nature of DRAM errors and the implications for system design. Proceedings of the 17th International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS), pp. 111–122

          Finding Memory is one of the leading hardware causes of crashes in data centres. A large share of the memory errors in the field are hard errors, meaning recurring ones, and not one-off glitches.

          In the tool Because memory errors come back, the test is worthwhile; the page thereby explains why a memory fault often produces quite different codes.

        9. Sridharan, V.; Liberty, D. (2012): A study of DRAM failures in the field. International Conference for High Performance Computing, Networking, Storage and Analysis (SC12), pp. 1–11

          Finding Eleven months of memory errors in a large compute cluster: permanent errors outnumber transient ones, and there are failures of whole rows or columns as well as errors in the circuitry on the board.

          In the tool Together with Schroeder 2009 and Hwang 2012, supports the ranking: if crashes accumulate, a memory fault becomes more likely.

        10. Meza, J.; Wu, Q.; Kumar, S.; Mutlu, O. (2015): Revisiting Memory Errors in Large-Scale Production Data Centers: Analysis and Modeling of New Trends from the Field. 45th Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN), pp. 415–426

          Finding Fourteen months of memory errors across Facebook’s entire server fleet: most errors came not from the memory chips but from the memory controller and memory channel.

          In the tool The memory test step advises testing modules individually and in another slot, rather than suspecting only the module.

        11. Schroeder, B.; Damouras, S.; Gill, P. (2010): Understanding latent sector errors and how to protect against them. ACM Transactions on Storage 6 (3), pp. 1–23

          Finding Unreadable sectors occur in the field in significant numbers and become more common with larger drives. Even a single one can cost data in systems without redundancy.

          In the tool For codes that point to read errors, the tool puts the backup before any repair.

        12. Ma, A.; Traylor, R.; Douglis, F. et al. (2015): RAIDShield: characterizing, monitoring, and proactively protecting against disk failures. ACM Transactions on Storage 11 (4), pp. 1–28

          Finding Failure data from around one million SATA drives over up to five years: the number of reallocated sectors is closely linked to an impending failure.

          In the tool The step “Check the hard drive or SSD” begins with the SMART values and refers to the hard drive check.

        13. Kalyanakrishnam, M.; Kalbarczyk, Z.; Iyer, R. (1999): Failure data analysis of a LAN of Windows NT based computers. Proceedings of the 18th IEEE Symposium on Reliable Distributed Systems (SRDS), pp. 178–187

          Finding Event logs of Windows NT mail servers over six months: most restarts had software causes, but the restart did not always solve the problem, and failures occurred in clusters.

          In the tool The restart comes first, but if you choose “Several times”, you get the full sequence of steps instead of advice to wait and see.

        14. Crameri, O.; Knezevic, N.; Kostic, D. et al. (2007): Staged deployment in Mirage, an integrated software upgrade testing and distribution system. Proceedings of the 21st ACM Symposium on Operating Systems Principles (SOSP), pp. 221–236

          Finding In a survey of 50 administrators, two thirds estimated that 5 to 10 per cent of the updates they installed cause problems; 18 per cent had already experienced severe outages caused by updates. Common reasons are broken dependencies and faulty updates.

          In the tool If the blue screen began after an update, “Uninstall the update” moves to the top of the steps.

        Standards, authorities and manufacturer data

        Mon–Fri 8:00–17:00

        Does the blue screen come back despite everything? Send us the copied result. We read the crash logs, remotely or on site in Rodewisch and throughout the Vogtland, and fix the cause.

        +49 3744 4399760[email protected]

        Mon–Fri 8:00–17:00 · Straße der Jugend 5, 08228 Rodewisch

        Have the crash analysed
        +49 3744 4399760Mon–Fri 8:00–17:00
        CallCallback

        PARTS LIST · SHEET 1/1

        What this page is made of

        Values measured at build time. With JavaScript, your browser measures them again live.

        ITEMPARTTYPESIZE
        01index.htmlDocument≈ 29 kB
        02styles.cssStylesheet≈ 20 kB
        03scripts.jsScript≈ 13 kB
        04archivo-variable.woff2Font≈ 58 kB
        05ibm-plex-mono.woff2Font≈ 30 kB
        06favicon.svgImage< 1 kB

        Advertising or analytics cookies: none. Without your consent, the page loads nothing from third-party servers apart from Cloudflare, our delivery service. Fonts, scripts and images are hosted by us. That is how we build your IT, too.

        OUTSIDE OFFICE HOURS

        Feel free to call anyway

        Our regular office hours are Monday to Friday, 8:00–17:00.

        In the evening and at weekends the number is forwarded to us, and often someone will pick up. We cannot promise that, though.

        If nobody answers, leave a short message on the voicemail: your name, your number and what it is about. We will call back as soon as we can, on the next working day at the latest.

        The call costs nothing. If someone has to start work straight away outside office hours, the surcharge in the price list applies.

        Call now+49 3744 4399760

        Prefer to write? Use the contact form or e-mail [email protected]. We will read it on the next working day.

        DIAGNOSIS · AT-K2

        What do you need help with?

        ↵ SELECTESC CLOSECALL: +49 3744 4399760

        Custom settings

        Which services may load?

        Strictly necessary

        The choice you make here and your settings for colour scheme and motion. These are kept as small entries on your device and are never sent to a server. They are needed for the site to work and therefore cannot be switched off.