Fix SYSTEM_THREAD_EXCEPTION_NOT_HANDLED BSOD: Driver Troubleshooting & Safe Mode

The SYSTEM_THREAD_EXCEPTION_NOT_HANDLED blue screen is one of those Windows errors that looks frightening but, in most cases, can be diagnosed and fixed without replacing hardware — provided you approach it methodically and understand what the stop code is telling you. This article breaks down what that stop code means, why graphics and other drivers are common culprits, and provides an ordered, practical troubleshooting plan with both conservative and advanced options so you can get a troubled PC back to stable operation.

Background / Overview​

The error shown as SYSTEM_THREAD_EXCEPTION_NOT_HANDLED (stop code 0x1000007E or variant) signals that a system thread raised an exception that the kernel’s exception handler didn’t catch. On modern Windows systems the crash screen will often show a filename (for example, nvlddmkm.sys, atikmdag.sys, igdkmd64.sys, or dxgmms2.sys) — and that filename is a huge diagnostic clue because it usually identifies the driver module that was running when the exception occurred. Graphics drivers are a particularly frequent cause of this specific BSOD, but the underlying cause can also be network drivers, storage drivers, or other kernel-mode components.
Why this matters: when a kernel-mode driver misbehaves it can violate memory access rules or raise an exception while running at an elevated privilege level. Windows stops to protect data integrity and hardware, producing the blue screen rather than letting the system continue in an inconsistent state.

---

How this BSOD typically affects your PC​

• Instant stop / reboot: Windows halts and usually restarts automatically, which can create boot loops if the root cause persists.
• Intermittent to constant crashes: Initially crashes may happen under load (gaming, video playback) and escalate into frequent or unbootable states.
• Data access risk: While the BSOD itself doesn’t usually destroy data, repeated crashes increase the risk of file-system corruption.
• Diagnostic hints: The crash screen or minidump files frequently list a filename which points to the suspicious driver — use that filename as your primary lead. Community diagnostics and professional guidance often begin by following that trace.

---

Common causes (what to look for first)​

• Outdated, corrupted, or incompatible device drivers — especially GPU/display drivers. These are the most common cause for SYSTEM_THREAD_EXCEPTION_NOT_HANDLED. If the crash names an NVIDIA, AMD, or Intel driver file, treat the GPU driver as the prime suspect.
• Faulty or misconfigured hardware — bad RAM, failing GPU, or unstable overclocks can throw kernel exceptions.
• Corrupted system files — damaged Windows components can make otherwise-good drivers crash the kernel.
• Storage/firmware issues — failing SSD/HDD or outdated SSD firmware can manifest as BSODs during heavy I/O.
• Third‑party system utilities and antivirus — low-level utilities (kernel-level AV, virtualization drivers, or custom system filters) can inject buggy drivers that crash Windows.
• Bad cable or display chain (edge case) — in recent community reports, mismatched or defective DisplayPort/HDMI cables and monitor adapters have been implicated in driver crashes on some GPU models; these reports are anecdotal but worth checking if other fixes fail. This type of claim should be treated as anecdotal and verified in your environment before assuming hardware replacement.

---

Immediate triage — what to do the moment you see the blue screen​

• Note the filename shown on the BSOD (if present). Writing it down makes subsequent troubleshooting faster.
• Capture a photo of the screen if it scrolls away on reboot. The filename and QR code matter.
• Try a clean reboot and check whether Windows restarts normally. If you hit a boot loop, move immediately to recovery options or Safe Mode.

If you can’t boot normally, proceed to Safe Mode as described below.

---

Boot into Safe Mode (first diagnostic environment)​

Safe Mode loads Windows with a minimal set of drivers and services and is the first step for isolating driver-related crashes.

• Force Windows into Automatic Repair by powering on and interrupting boot 2–3 times (power-button shutdown during POST or Windows splash). When Windows boots to Recovery, choose:
  Troubleshoot → Advanced options → Startup Settings → Restart → press 4 (Enable Safe Mode) or 5 (Safe Mode with Networking).
• Alternatively, from a working Windows session: Settings → Update & Security → Recovery → Advanced startup → Restart now → Troubleshoot → Advanced options → Startup Settings.

If the BSOD does not appear in Safe Mode, the problem is almost certainly caused by a third‑party driver or utility (not a core Windows system file). Community troubleshooting workflows follow this logic reliably.

---

Step-by-step troubleshooting: from least invasive to most invasive​

1. Identify the problematic driver from the crash hint​

• If the BSOD lists nvlddmkm.sys, investigate NVIDIA drivers.
• If it lists atikmdag.sys or atikmdag64.sys, investigate AMD drivers.
• If igdkmd64.sys is listed, check Intel integrated graphics drivers.
• If dxgmms2.sys appears, the issue is often tied to DirectX/graphics stack or hardware acceleration. (partitionwizard.com, drivereasy.com)

These filenames aren’t definitive proof — they indicate the module that failed while handling an operation. Use them as your first lead.

2. Use Device Manager to update or roll back the driver​

• Press Win + X → Device Manager.
• Expand Display adapters (or the category indicated by the BSOD filename).
• Right-click the device → Properties → Driver tab.
• If the crash started after a driver update, click Roll Back Driver (if available).
• Otherwise, click Update driver → Search automatically, or use the manufacturer’s download.

If Windows Update or Device Manager doesn’t find a good driver, download and install the driver directly from the vendor: NVIDIA, AMD, Intel, or your PC maker. Installing the official vendor package (not a generic Windows driver) is often the fix. Tom’s Hardware and Microsoft community guidance both point to driver updates/rollbacks as the first practical step. (tomshardware.com, answers.microsoft.com)

3. Cleanly remove and reinstall display drivers (use DDU for graphics stacks)​

• Boot to Safe Mode.
• Use Display Driver Uninstaller (DDU) to completely remove GPU drivers and registry remnants.
• Reboot and install the latest stable driver from the GPU maker.

DDU is widely recommended in community troubleshooting because it prevents leftover driver components from provoking repeat crashes.

4. Run SFC and DISM to repair system files​

Run an elevated Command Prompt and execute:

• DISM /Online /Cleanup-Image /RestoreHealth
• sfc /scannow

Microsoft recommends running DISM first (it replenishes the repair store), then SFC. These built-in tools will repair corrupted system files that could be causing or contributing to the crashes. If SFC reports unrepairable files, follow Microsoft’s guidance for manual replacement or offline repair. (support.microsoft.com, learn.microsoft.com)

5. Check disk health and filesystem integrity​

• Run chkdsk C: /f /r from an elevated Command Prompt and schedule on reboot if necessary.
• Use vendor SSD utilities (Samsung Magician, Crucial Storage Executive, etc.) to check SMART status and update firmware.

Bad sectors, failing controllers, or outdated SSD firmware have been linked to kernel crashes in some cases — particularly when heavy I/O occurs. Community troubleshooting frequently includes chkdsk and disk utility checks early in the process.

6. Disable overclocking and reset BIOS defaults​

Unstable overclocks on CPU, GPU, or RAM can trigger kernel exceptions that look like driver faults. Reset BIOS/UEFI to default settings and disable XMP profiles or any manual tuning before testing.

7. Use System Restore to roll back recent changes​

If crashes began after a recent update, driver install, or new software, restore Windows to a previous restore point. System Restore doesn’t affect personal files but will remove drivers and updates applied after the selected point.

8. Advanced: Driver Verifier and minidump analysis​

• Enable Driver Verifier to stress drivers and force the faulty driver to reveal itself (Driver Verifier may itself cause BSODs while testing — only use this for diagnosis).
• Collect minidump files from C:\Windows\Minidump and analyze them in WinDbg or BlueScreenView to identify the failing module more precisely.

Forums and advanced guides recommend these techniques when standard updates and clean installs fail.

---

Fixes for specific driver modules (practical tips)​

nvlddmkm.sys (NVIDIA)​

• Use DDU in Safe Mode to fully remove NVIDIA drivers, then install a stable version from NVIDIA’s site.
• For laptops, prefer vendor-supplied drivers (OEMs often tailor drivers to the system).
• Check power delivery, PCIe seating, and cables — hardware or power issues can imitate driver faults. Anecdotal reports show faulty power cabling or GPU sag sometimes correlated with nvlddmkm.sys crashes (treat these as field reports that require verification). (tomshardware.com, reddit.com)

atikmdag.sys (AMD)​

• Roll back to a prior working AMD driver if a recent update caused the crash. Use DDU for clean uninstall.
• Check for conflicts with third-party overlays (Discord/Steam/GeForce Experience overlay).

dxgmms2.sys (DirectX / Windows graphics subsystem)​

• Update GPU drivers and Windows.
• Consider disabling hardware acceleration in apps (office suites, browsers) if crashes happen during video playback or browser use.
• Adjust or increase the TdrDelay registry value if the GPU occasionally hangs but recovers; this is a workaround, not a cure. DriverEasy and PartitionWizard document registry and acceleration tweaks commonly used to mitigate dxgmms2.sys crashes. Use caution and back up the registry before editing. (drivereasy.com, partitionwizard.com)

---

When to suspect hardware and what to test​

• RAM: Run Windows Memory Diagnostic or MemTest86. Memory errors can produce cryptic BSODs.
• GPU: Test another GPU if available, or test your GPU in another system. Check for physical damage and ensure the card is firmly seated and powered.
• PSU: Unstable power can produce odd driver errors — verify rails and connectors, especially with power-hungry GPUs.
• Storage: Check SMART, run surface tests, and test with a spare drive if possible.

If multiple components fail stress tests, consider professional diagnostics or RMA paths under warranty.

---

Advanced repair and last-resort options​

• In-place repair (Windows repair install / “Repair upgrade”)
• Reinstalls Windows while preserving apps, settings, and files. This often fixes unresolved driver or component corruption without data loss.
• Clean install of Windows
• Back up everything first. Delete partitions and perform a clean Custom install. After install, install chipset and display drivers from the vendor before restoring other software.
• Replace failing hardware
• Only after rigorous testing should you assume hardware fault. Community guides warn against premature replacement: run memory, GPU, disk, and PSU checks to confirm.

---

Prevention and long-term hygiene​

• Keep drivers updated on a regular schedule, but favor stable vendor releases over latest-beta packages for production machines.
• Maintain regular system backups and restore points prior to major driver or OS updates.
• Use reputable antivirus and avoid low-level system utilities of uncertain provenance.
• Monitor storage and memory health periodically with vendor tools and S.M.A.R.T. monitors.
• When installing new hardware, uninstall old drivers completely (use DDU for GPU swaps) before installing replacements.

Community threads emphasize disciplined driver management and making restore points before changes as the most effective prevention strategy.

---

Critical analysis — strengths of common approaches and their limits​

• Driver update / rollback (Strength): Quick and effective for the most common root cause (graphics drivers). It’s non-destructive and reversible when performed correctly.
• SFC/DISM (Strength): Repairs corrupt system files without reinstalling Windows and is officially supported by Microsoft; usually a reliable early step.
• Safe Mode & DDU (Strength): Safe Mode isolates the driver layer; DDU removes driver residues that often cause repeat failures. Communities widely endorse this sequence.
• Driver Verifier and dump analysis (Strength): These provide objective, technical evidence of the root cause and can avoid guesswork in complex cases.

Limitations and risks:

• Blindly reinstalling the “latest” driver (Risk): New drivers can introduce regressions. For production machines prefer known good driver versions rather than the bleeding edge.
• Registry or firmware edits (Risk): Adjustments like TdrDelay or registry toggles can mitigate symptoms but don’t fix root causes; they may also hide failing hardware issues.
• Driver Verifier (Risk): It intentionally stresses drivers and will likely produce more BSODs during diagnosis — only enable it for targeted troubleshooting and be prepared to revert changes.
• Anecdotal fixes (Risk): Community reports of odd causes (DP cable compatibility, GPU sag, or specific PSU wiring) can be correct in some cases, but these are environment-specific and hard to generalize. Treat them as testable hypotheses rather than established solutions.

---

Practical checklist — a compact sequence to follow​

• Note the BSOD filename and time; photograph the screen.
• Boot to Safe Mode. If stable, suspect drivers.
• In Safe Mode: run SFC/DISM, and scan disk health (chkdsk).
• Remove and reinstall suspect driver(s) with DDU for GPUs.
• Test by rolling back to previously-working drivers if the latest fails.
• Run memory and disk diagnostics (MemTest86, vendor SSD tools).
• If unresolved, collect minidumps and run WinDbg or get help with dumps.
• Consider in-place repair, then clean install as a last resort.

This ordered checklist reflects best-practice community workflows and Microsoft’s recommended repair tools.

---

When to get professional or manufacturer help​

• If diagnostics show hardware failure (RAM errors, SMART reports of failing sectors, or GPU artifacting), contact the device manufacturer or your PC vendor for warranty support.
• If the system is under warranty, vendor technical support can run hardware-specific diagnostics that aren’t available publicly.
• Pay a certified technician if you lack backup or cannot risk data loss — an expert can image drives and perform safe recovery before attempting destructive repairs.

Community advice and manufacturer support channels both recommend warranty/tech support once hardware faults become likely.

---

Final thoughts and key takeaways​

• Most SYSTEM_THREAD_EXCEPTION_NOT_HANDLED crashes are driver-related, with GPU/display drivers appearing most often. Start with Safe Mode, driver rollback/clean reinstall, and SFC/DISM. (tomshardware.com, support.microsoft.com)
• Work methodically: document the BSOD filename, try non-destructive fixes first, then escalate to more invasive options.
• Use logs and dumps: Minidumps and Event Viewer are critical for accurate diagnosis; they separate guesswork from evidence.
• Watch out for anecdotal fixes: Reports of cables, sagging GPUs, or power-cabling oddities do crop up in real-world troubleshooting — but validate those in your system before assuming a hardware swap.
• Back up before making big changes: in-place repairs and clean installs are powerful but always carry risk without a recent backup.

If the BSOD returns repeatedly after following the recommended steps, the odds shift from a flaky driver to possible hardware or deep system corruption — at that point rely on hardware diagnostics and manufacturer support rather than repeated software tweaks. Communities and support documentation converge on careful, evidence-based troubleshooting as the fastest path to a durable fix.

---

This guide is intended to be a practical, stepwise reference for resolving SYSTEM_THREAD_EXCEPTION_NOT_HANDLED stops. It combines Microsoft’s official troubleshooting tools with field-tested community procedures to help you identify the real culprit — and to fix it without unnecessary parts swaps or data loss.

---

Source: 9meters System Thread Exception Not Handled - 9meters