How to fix "The description for Event ID 153 from source nvlddmkm cannot be found" in Windows?

Event ID 153 from source nvlddmkm is the NVIDIA display driver writing down that the graphics card stopped answering and was reset. The sentence about a missing or corrupted component is Windows boilerplate, not the fault. If your games crash or the screen goes black at the same minute, start by reading the driver's own line, then try an older driver.

Facts checked on 15 October 2026. Steps written for Windows 11 (25H2 and 26H2) and Windows 10 22H2 with an NVIDIA GeForce graphics card.

What "The description for Event ID 153 from source nvlddmkm cannot be found" means

nvlddmkm is the kernel part of the NVIDIA graphics driver, the file nvlddmkm.sys. It writes its events to the System log with a number and a few raw values, but it ships no readable sentence for number 153. Event Viewer therefore prints its standard text for a missing description and shows the raw values below it.

The full entry as it appears in Event Viewer (the second line differs between PCs)

The description for Event ID 153 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.

If the event originated on another computer, the display information had to be saved with the event.

The following information was included with the event:

\Device\Video3
Error occurred on GPUID: 100

The message resource is present but the message was not found in the message table

• "Cannot be found" and "installation is corrupted" are Windows guesses. The same sentence appears for any program that has no text for an event number. It does not mean your NVIDIA driver is damaged.
• "The message resource is present but the message was not found" says the opposite: Windows found the driver's message file, and number 153 is simply not in it.
• The useful part is the line after "The following information was included with the event". For Event ID 153 it names the device (\Device\Video3) and what happened on the GPU.
• Event ID 153 is not a Windows error code. The number belongs to the nvlddmkm source only. An Event ID 153 from another source is unrelated.

Error occurred, Reset TDR, Restarting TDR: what the line under Event ID 153 tells you

TDR stands for Timeout Detection and Recovery. Microsoft documents it: when the graphics card cannot finish or interrupt a task within two seconds, Windows treats it as frozen, resets the driver and the card, and redraws the desktop. The only visible sign of a successful recovery is a screen flicker and the note "Display driver stopped responding and has recovered".

• Users describe it the same way. One report on Microsoft Q&A: the PC freezes for about three seconds, goes black for a second, then recovers. That is a TDR.
• Older games may stay black. Microsoft notes that some legacy DirectX programs render only black after a recovery and must be restarted, which looks like a crash.
• Too many resets end in a blue screen. By default Windows stops with a bug check when five or more GPU hangs occur within one minute.
• Video3 and GPUID 100 are not a fault code. They differ between PCs, and we could not confirm from Microsoft or NVIDIA what they encode.

Why nvlddmkm logs Event ID 153: the causes in the order to rule them out

Nobody publishes how often each cause occurs, so the order below follows two things: what users with this exact event most often report as their cure, and NVIDIA's own checklist for crashes and freezes, cheapest step first.

How to fix the crashes behind Event ID 153 from source nvlddmkm

Always start with Fix 1, which tells you which of the others you need. The rest go from free and reversible to opening the PC case. Stop as soon as a full gaming session passes without a new Event ID 153.

Fix 1. List the Event ID 153 entries and the events around them

10 minutesNo riskEveryone who sees Event ID 153, before changing anything.

One crash often leaves several 153 entries in the same second. Grouping them into incidents, and seeing what Windows logged next to them, tells you whether you face a driver reset, a power problem on the PCIe link, or nothing at all.

1. Right-click Start and select Terminal (Admin). On Windows 10, select Windows PowerShell (Admin).
2. Paste this command and press Enter. It lists the last 30 events with ID 153 from the NVIDIA driver and the line the driver stored.

   Get-WinEvent -FilterHashtable @{ LogName='System'; ProviderName='nvlddmkm'; Id=153 } -MaxEvents 30 |
     Select-Object TimeCreated, @{ n='Line'; e={ $_.Properties.Value -join ' ' } } |
     Format-Table -Wrap

3. Group the results by time. Entries within a few seconds of each other are one incident. Note the date, the time and whether the line says Error occurred or Reset TDR occurred.
4. Check for hardware warnings from the PCI Express bus at the same times:

   Get-WinEvent -FilterHashtable @{ LogName='System'; ProviderName='Microsoft-Windows-WHEA-Logger' } -MaxEvents 30 |
     Format-Table TimeCreated, Id, Message -Wrap

5. Press Win + R, type perfmon /rel and press Enter. On the days of your incidents, look for critical entries that name LiveKernelEvent with the code 141 or 117: they are recovered GPU timeouts, not blue screens.
6. Open C:\Windows\LiveKernelReports in File Explorer. Dump files with the dates of your incidents confirm the driver resets.

Fix 2. Remove the driver in Safe Mode and install the previous version

30 minutesLow riskThe incidents started after a driver update, including one installed automatically by the NVIDIA app or Windows Update.

Going back to an older driver is the cure users report most often for this event. On Microsoft Q&A, the person who asked fixed it with a February driver, and a laptop owner fixed artifacts and restarts by returning from 581.29 to 572.70. NVIDIA also notes that a newer driver can stress the card harder, so a once-stable setup can start to fail.

1. On NVIDIA's driver page, find your card and download the version you used before the problems began. If you do not know it, take the one released before your current driver.
2. Download Display Driver Uninstaller (DDU), the free tool NVIDIA's support article names for removing old driver installs. NVIDIA calls it a tool for advanced users and it is not an NVIDIA product. Extract the archive.
3. Disconnect from the internet, so Windows does not install a driver on its own after the cleanup.
4. Start Windows in Safe Mode: Start Windows or a Mac in Safe Mode.
5. Open Display Driver Uninstaller.exe, select GPU and NVIDIA in the drop-down lists on the right, and click Clean and restart.
6. After the restart, run the older driver installer, select Custom (Advanced) and tick Perform a clean installation.
7. Restart, reconnect to the internet and play the game that crashed for at least as long as it used to take to fail.

• Quick alternative. If Device Manager still holds the previous driver, Roll Back Driver on the card's Driver tab does the same in a few clicks: Update, roll back or reinstall a driver.
• Keep the older version. Turn off automatic driver updates in the NVIDIA app while you test, and check the Driver Version on the Driver tab after a few days.
• Laptops. If the NVIDIA driver keeps failing on a laptop, try the graphics driver from the laptop maker's support page.

Fix 3. Run the graphics card, the processor and the memory at default speeds

15 minutesLow riskAny part is overclocked, a tuning tool such as a GPU clock utility runs at start-up, the card is a factory-overclocked model, or an XMP or EXPO memory profile is on.

A GPU timeout is what an unstable clock looks like to Windows. NVIDIA's checklist says newer drivers use the card more fully, so an overclock that was stable on an old driver may no longer hold, and asks you to set every overclocked part, including the processor and memory, back to default.

1. Open your GPU tuning tool, reset clocks, voltage and fan curves to default, and stop it from starting with Windows: Stop apps from opening at startup.
2. Open the NVIDIA app, go to System and then Advanced, and turn on Debug Mode. It runs the card at NVIDIA's reference clocks, below a factory overclock.
3. Restart and enter the firmware setup (BIOS or UEFI). The key is shown at start-up, often Del, F2 or F10.
4. Set the memory profile (XMP or EXPO) to Default or Disabled, and reset any manual processor overclock.
5. Save, start Windows and play the game that crashed.

• Stable now: one of the settings was the cause. Turn them back on one at a time, test after each, and keep what is stable.
• Stable only with Debug Mode: the card cannot hold its factory clocks any more. That is a warranty matter for the card maker.
• Still crashing: clocks are ruled out. Keep the defaults while you continue, then turn Debug Mode off when you are done.

Fix 4. Install the chipset driver and check the BIOS version

20 minutesRead the warningA self-built PC, a new motherboard, or a fresh Windows install where the motherboard maker's drivers were never installed.

NVIDIA's checklist singles this out for custom-built PCs: the basic drivers Windows installs for the motherboard are fine for light use, but under intensive programs a PC can run better with the chipset driver from the chipset maker. The chipset controls the PCIe link the graphics card talks through.

1. Press Win + S, type System Information and open it.
2. Note BaseBoard Manufacturer, BaseBoard Product and BIOS Version/Date.
3. On the motherboard maker's support page for that exact board, download the latest chipset driver for Windows 11 or Windows 10, install it and restart.
4. On the same page, compare the newest BIOS version with yours. If the list of changes mentions stability, PCIe or graphics card compatibility, update it with the method the maker describes for your board.
5. Load the BIOS defaults after the update, so no old overclock comes back.

Fix 5. Turn off in-game overlays and close background apps while you test

10 minutesNo riskThe crash comes when you Alt + Tab, open an overlay, or when a recorder, a chat app or a performance monitor runs over the game.

Overlays draw on top of the game through the same graphics card. Users report that turning them off stopped the 153 crashes, and NVIDIA's checklist asks you to close programs the game does not need before launching it. We found no vendor statement that overlays cause this event, so treat this as a test.

1. In Steam, open Settings > In Game and turn off Enable the Steam Overlay while in-game.
2. In Discord, open User Settings, find the in-game overlay option and turn it off.
3. In the NVIDIA app, open Settings and turn off the overlay.
4. Close recording, RGB, monitoring and tuning tools from the notification area.
5. Play the game that crashed. If it now runs, turn overlays back on one at a time to find the one that matters.

Fix 6. Check power, the PCIe slot and temperatures

30 to 40 minutesLow riskCrashes come under heavy load or after minutes of play, WHEA-Logger warnings appeared in Fix 1, the card sits on a riser cable, or the fans are loud.

When the card loses power or its PCIe link for a moment, it stops answering and Windows logs a reset. NVIDIA's checklist names dust, airflow and the power supply, and notes that wattage alone does not decide whether a power supply suits a card. One Event ID 153 report on Microsoft Q&A came with constant WHEA-Logger warnings from a PCI Express Root Port.

1. Press Ctrl + Shift + Esc, open Performance and select the GPU. Note its temperature while the game runs in a window, up to the moment of a crash.
2. Shut down, switch off the power supply and unplug the PC.
3. Press the graphics card firmly into its slot and check its bracket screws.
4. Unplug and firmly reconnect each power cable of the card, at the card and at the power supply. Use separate cables from the power supply when the card has more than one input and you have spare cables.
5. If the card sits on a riser cable, plug it straight into the motherboard for the test.
6. Blow the dust out of the card's cooler and the case fans, close the case and test again.

• Temperature. We give no limit, as it depends on the card. A temperature that climbs to the same value before every crash points to cooling.
• Placement. NVIDIA warns that a PC in a closed desk compartment can overheat. Test with the side panel off.
• The decisive test. If you can borrow a power supply or put the card into another PC, do it. That tells you more than any setting.

Fix 7. Test the system memory

30 minutes or moreNo riskCrashes also happen outside games, or the PC became stable in Fix 3 only after XMP or EXPO was turned off.

NVIDIA notes that faulty memory, processor or motherboard can show symptoms only in intensive games and look like a graphics card problem. The driver keeps its data in system memory, so a bad module can make nvlddmkm fail on a healthy card.

Opens Windows Memory Diagnostic

mdsched.exe

1. Open Windows Memory Diagnostic. Press Windows + R, type mdsched.exe and select OK. You can also search for Windows Memory Diagnostic and open it from the results.
2. Choose when to test. Select Restart now and check for problems (recommended) to test right away. Select Check for problems the next time I start my computer if you want to finish your work first: the test then runs at the next start.
3. Let the test run. The PC restarts into a blue screen that shows the progress of the test. You do not have to do anything. When the test ends, the PC restarts by itself and Windows loads.
4. Look for the notification. Sign in. After a short while Windows shows a notification with the result, also when no errors were found. If you missed it, read the result in Event Viewer.
5. Open the System log. Press Windows + R, type eventvwr and select OK. In Event Viewer, expand Windows Logs and select System.
6. Find the result. Right-click System and select Find. Type MemoryDiagnostics-Results and press Enter. Select the event that is found: its text says whether the test detected errors.

These steps were written for Windows 11 25H2 and 26H2, Windows 10 22H2. The full page with the other versions and what each result means: How to test your RAM with Windows Memory Diagnostic

Errors are a clear result: test one module at a time to find the bad one. No errors is weaker evidence, because a short test can miss faults that show only when the memory is warm. If the PC is stable only with the memory profile off, keep it off or pick a slower profile.

If Event ID 153 keeps coming back: what to collect and when it is the hardware

After the fixes you have a PC at default speeds, a cleanly installed driver in two versions, current chipset drivers, checked cables and tested memory. If 153 entries still come with crashes, software is unlikely. NVIDIA's checklist gives the deciding test: try the card in another computer.

• Signs of hardware: crashes at default speeds with every driver and in every 3D program, WHEA-Logger warnings, and a PC that restarts with no dump.
• Signs of software: one game only, a clear start date that matches a driver or game update, and stability after a rollback.
• What to send to support: your incident list from Fix 1 with the driver's lines, the driver versions you tried, the card and power supply models, and the files from C:\Windows\LiveKernelReports.
• Is a repair install of Windows worth it? Only when the PC also has problems outside graphics. It does not touch the NVIDIA driver's missing description, and a clean driver install already replaces the driver files.

How to keep Event ID 153 from coming back

Frequently asked questions

What is Event ID 153 in Windows?

Event ID 153 is not a Windows error code. Every program that writes to the event log numbers its own events, and the number only means something together with the source. With the source nvlddmkm it is an event of the NVIDIA graphics driver, and the line inside it usually reads "Error occurred on GPUID: 100" or "Reset TDR occurred on GPUID:100", which points to a graphics card reset after a timeout. An Event ID 153 from another source, such as a disk or network driver, is a different event and is not covered here.

Is my NVIDIA driver really not installed or corrupted?

No, the sentence does not prove that. Event Viewer shows it for any event whose number has no text in the program's message file. The last line of the entry, "The message resource is present but the message was not found in the message table", confirms that the NVIDIA driver's message file was found. A driver that is really missing shows elsewhere: the Driver tab of the card in Device Manager then names no NVIDIA driver, and games do not start at all. A clean reinstall can still help with the crashes, but not with the wording.

Does Event ID 153 mean my graphics card is dying?

Not by itself. The same event comes from one faulty driver version, an overclock, a memory profile, a loose power cable, dust or a weak power supply, and users most often report that an older driver cured it. A failing card becomes likely only when the crashes continue at default speeds, with more than one cleanly installed driver, in every 3D program, and above all when the card also fails in another PC. That swap test is what NVIDIA itself recommends.

Some answers say Event ID 153 means disk I/O delays. Is that right?

We found no Microsoft or NVIDIA source for it. An answer on Microsoft Q&A says Event ID 153 from nvlddmkm often appears with disk I/O delays, but that answer comes from a community member, not from Microsoft or NVIDIA. Disk drivers use their own sources, and a disk event 153 is unrelated to nvlddmkm. The lines this driver logs, Error occurred and Reset TDR occurred on a GPUID, describe the graphics card. Check the disk only if the System log also shows disk errors from a disk source.

What is the difference between nvlddmkm Event ID 14 and 153?

Both come from the NVIDIA driver and have no description text. Event ID 14 carries a line of register names and hexadecimal values after the device name, while Event ID 153 carries a short text such as "Reset TDR occurred on GPUID:100". Users often find both in the same second, which suggests one incident logged in several entries. NVIDIA does not document either number, so treat them as one problem and work through the same fixes. Our Event ID 0 guide covers IDs 13 and 14 in more detail.

Will increasing TdrDelay in the registry fix Event ID 153?

No. TdrDelay sets how long Windows waits before it declares the graphics card frozen, and the default is two seconds. A longer wait does not make the card answer: you see a frozen picture for longer before the same reset. Microsoft's documentation says the TDR registry values are meant for testing and debugging drivers. If you already added such values, remove them after backing up the key, as described in our Event ID 0 nvlddmkm guide.

Why does Event ID 153 appear even when the PC is idle or browsing?

Users report it during light load too: one laptop owner on Microsoft Q&A saw it mostly while browsing and in office programs, and even when idle, but not in games. In such reports the cure was again an older driver. Light load means the card changes its clock speeds and power states often, so a driver or power-state fault can surface there. Start with Fix 2, then check that no tuning tool changes clocks in the background.

Can I hide or stop Event ID 153 from being logged?

You can filter it out of the Event Viewer view, but you should not try to stop it being logged. The event is the driver telling you that the card was reset. Hiding it changes nothing about the freezes or crashes that come with it. If the entries appear without any symptom at all, you can simply ignore them: a recovered reset that you never notice does no harm. If they come with crashes, fix the cause with the methods on this page.

The message is in Russian or another language. Is this the same problem?

Yes. Windows translates its boilerplate into the system language, so a Russian Windows shows a sentence starting with the words for "the description for event ID 153 from source nvlddmkm cannot be found". The driver's own line, such as "Reset TDR occurred on GPUID:100", usually stays in English because it comes from the NVIDIA driver. The meaning and the fixes are the same as on this page.