PC Randomly Freezes: Sort the Freeze, Then Fix It (2026)

By LK Wood IV · 2026-10-02 · ~10 min read · St. Louis County, MO

Two seconds. By default, that is how long Windows waits on a graphics card before it decides the card has hung and resets it. Storage has a timer of its own. When either runs out, Windows writes a record. It writes another at the next start if the last session did not shut down cleanly.

Read those records first and the list of suspects gets short.

A PC that restarts or switches off by itself is a different fault. My PSU failure symptoms guide starts there. A game that hitches while everything else keeps running is stutter, which has fixes of its own.

While it is frozen, check what still responds

All six checks together take under a minute. Work down the list and note the first one that gets a response.

1. Move the mouse. A cursor that still moves means Windows is running and something it depends on has stalled.
2. Press Caps Lock and watch its light. This is Microsoft’s own test. If the light no longer changes, the computer may be in what Microsoft calls a hard hang.
3. Press Ctrl+Shift+Esc. If Task Manager opens, look for a program marked Not responding, select it and choose End task.
4. Press the Windows key with Ctrl, Shift and B. Microsoft documents it as the shortcut that resets the graphics driver, and a beep or a flicker means it ran.
5. Press Ctrl+Alt+Del. Microsoft’s freeze guide uses it as a test of whether Windows still answers. If the security screen appears, sign out or open Task Manager from there.
6. If nothing answers, note the time, then hold the power button until the PC turns off. The log is sorted by time and you will want it.

Match what you saw to a kind of freeze

After the restart, read what Windows wrote down

Start with the quick view. Press the Windows key with R, type perfmon /rel and press Enter, which Microsoft’s command reference lists as the switch that starts Reliability Monitor, then pick the day of the freeze. A session you ended with the power button shows as an entry reading “Windows was not properly shut down.” Note its time.

Then open the full record. Press the Windows key with R, type eventvwr.msc, and in the tree on the left go to Windows Logs, then System.

The record of the restart

Find the Critical record from the source Kernel-Power with Event ID 41, stamped at the moment the PC came back up. On my desktop, Windows 11 Home 25H2 build 26200.9457, its text reads “The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly.”

That sentence names three possibilities and picks none. Microsoft’s guide to the event agrees. The record is written at the next start, after the fact, and by itself it may not hold enough to say what happened. Two fields in its details narrow it.

• BugcheckCode is zero when Windows recorded no stop error. Any other value is the code of a stop error, written in decimal.
• PowerButtonTimestamp is nonzero when Windows recorded the power button being held. It can stay at zero after a hang, because a hung system may have been unable to write anything.

A BugcheckCode above zero turns a freeze into a stop code with a name. Five of them have a page here, and the name is the one Windows prints in capitals with underscores, as in MEMORY_MANAGEMENT.

For any other number, switch Windows Calculator to Programmer mode, enter the value as decimal and read it as hex, then look it up in Microsoft’s Bug Check Code Reference.

The minutes before it

A second record, Event ID 6008 from the source EventLog, prints the time the last session died: “The previous system shutdown at” a time and date “was unexpected.” Microsoft’s guide says to look for anything suspicious before that time in both the System and the Application log, and these are the records I would look for, with the wording from my desktop.

Microsoft’s driver documentation describes WHEA-Logger as the provider meant for desktop users, with a readable message carrying the main details. On my desktop those messages name the part: Processor Core, Memory, or a device given by its bus address and vendor ID. Event ID 129 is filed under the storage driver’s own name. On my desktop the registered ones are stornvme for NVMe drives and storahci for SATA.

A hard hang with nothing else in the log

Kernel-Power 41 with a BugcheckCode of zero and nothing logged before it is what a hard hang leaves, because Windows stopped before it could write anything. Microsoft’s guide says these symptoms typically indicate a hardware problem. Its list of checks starts with four — overclocking, memory, the power supply and overheating — and I would take them in order of what each costs to test.

1. Undo the last change. On a Tom’s Hardware Forum thread from October 2022, the owner of a Ryzen 5 5600X and RX 6600 had hard freezes in games, with the keyboard’s lock lights no longer changing and a hard reset the only way out. Their only recent change was a new 240 Hz monitor. They reported a week without a freeze after turning FreeSync off on it. A second member with a Ryzen 9 3950X and RTX 3080 reported two weeks without one after a Windows update and a UEFI update. Neither replaced a part.
2. Put everything back to stock. Turn off the memory profile first. Intel describes XMP as overclocking the memory, and AMD’s full name for EXPO is Extended Profiles for Overclocking, so both fall under Microsoft’s first check. Then remove any overclock or undervolt on the processor and the graphics card. My memory management guide covers why a rated memory profile can be unstable on healthy sticks and how long to run at default before you trust the result.
3. Test the memory. Press the Windows key with R, type mdsched.exe, and choose “Restart now and check for problems (recommended).” The result lands back in the System log from the source MemoryDiagnostics-Results: Event ID 1101 or 1201 when it found no errors, 1102 or 1202 when it “detected hardware errors.” A clean quick test is weak evidence, and the same guide covers the overnight test that settles it.
4. Check heat under the load that triggers the freeze. How to check CPU temp covers the tools, and CPU overheating sorts a real fault from a chip running hot by design.
5. Suspect power last, because it is the hardest to test. My PSU failure symptoms guide has the tests. Its main sign is a shutdown or restart under load, and it lists freezes among the signs of unstable power.

WHEA-Logger records change the order. If the text names Memory or Processor Core, skip to steps 2 and 3.

A display hang

Microsoft’s display driver documentation says the frozen look often comes from the graphics card being busy, typically during game play, so that it stops updating the screen. By default Windows gives it two seconds. Then it resets the driver, and the only things you see are a flicker and a message that the display driver stopped responding and has recovered. Display 4101 is that reset in the log.

One recovered reset is Windows working. A run of them is a fault, and so is a BugcheckCode of 278, the stop code Windows raises when a reset fails or resets pile up. My GPU driver crash guide has the order, which starts with a clean driver install and every overclock and undervolt back to stock.

Stalls that recover

Open Task Manager with Ctrl+Shift+Esc the next time it happens, or straight after, and look at which column is full. Then match it to the log.

The drive stopped answering

Windows times every request it sends to a drive. Microsoft’s write-up of Event ID 129 describes what happens when that timer runs out: the storage driver decides the device has stopped responding, resets it and retries what was waiting. Anything that needs the drive has to wait for that, and I take that wait to be the pause you feel. The write-up is about storage area networks. The report below shows the same event on a laptop.

One Windows 11 laptop owner on Microsoft Q&A described it in March 2025. The cursor lagged, the machine stopped answering, Task Manager showed the disk at 100 percent, and every freeze lined up with a 129 followed by a run of 153s. It had gone on for over a year. ASUS had the laptop twice and replaced the motherboard. System file repairs, a factory reset, a BIOS update and a driver reinstall did not end it either, and their last post said they were looking at a new SSD.

So read the drive’s health first. My SSD health check reads an NVMe drive from Settings with no download and names the tool a SATA drive needs. If it reports a warning, copy your files off first. Microsoft’s write-up of Event 153 says an occasional timeout can be normal and frequent retries mean a storage problem to correct. A clean health report with 129s still arriving moves my suspicion to the connection: on a SATA drive I would swap the data cable, and on an M.2 drive I would reseat it.

Windows ran low on memory

Event ID 2004 names the three programs that were using the most virtual memory when Windows ran low. Look for a repeat. If the same program leads every time, close it, update it or replace it. I would add RAM only when the three are programs you need open together.

One program, or a service behind it

Application Hang 1002 names the program. Update it or reinstall it. When the log names nothing and the stalls continue, a background service is my next suspect. Microsoft’s clean boot procedure turns off every service that is not Microsoft’s, then has you turn them back on half at a time until the stall returns.

Damaged system files

Microsoft lists freezing among the symptoms of corrupted system files. Its repair takes two commands in an administrator Command Prompt, in this order.

DISM.exe /Online /Cleanup-image /Restorehealth
sfc /scannow

They repair Windows files and nothing else. On a drive that is still logging 129s, I would expect the damage to return.

Freezes at wake or soon after power-on

A hang on the way into sleep or out of it makes me suspect a driver first, then the firmware. Check the BugcheckCode for 159. That is stop code 0x9F, DRIVER_POWER_STATE_FAILURE, which Microsoft defines as a driver in an inconsistent or invalid power state, and it is the number in Microsoft’s own Event ID 41 example. I would update the chipset and graphics drivers from the maker’s site, then the BIOS. Until the freezes stop, use Shut down in place of Sleep.

A freeze soon after power-on that never follows a restart makes me suspect fast startup. With it on, Microsoft’s documentation says Windows saves the kernel and its loaded drivers to the hibernation file at shutdown and loads that file at the next start. I read that as a driver in a bad state at shutdown coming back in the same state at power-on. Restart is different. The option’s own description says restart is not affected.

To turn it off, press the Windows key with R, type powercfg.cpl and press Enter. Select “Choose what the power buttons do,” then “Change settings that are currently unavailable,” and clear “Turn on fast startup (recommended).” Then give it a week of cold starts.

What I would leave for last

A reset or a reinstall. A Microsoft Q&A thread from January 2025, marked as the same question by more than 40 people, is about a freeze of three to four seconds that kept returning, and it ends with a reset that the last post says solved it. Nobody in the thread asked what the System log showed. The cause was never named. The laptop owner with the 129s had a factory reset too, and the freezes came back.

Read the log first. If it holds nothing but Kernel-Power 41 after a week at stock settings and a clean overnight memory test, the power supply and the motherboard are what is left — and I would borrow a known-good power supply before buying either.