You're gaming or watching a video, and suddenly your screen goes blue. VIDEO_TDR_FAILURE. Your system restarts. Maybe it keeps restarting. And every time, that error message mentions atikmpag.sys. I've seen this problem hundreds of times through remote support, and I'm going to walk you through what's actually happening and how to fix it properly.
TL;DR
VIDEO_TDR_FAILURE atikmpag.sys crashes happen when AMD graphics drivers timeout or GPU hardware fails. Fix it by updating AMD drivers from AMD.com directly, disabling PCI Express power management, and cleaning dust from your GPU. Most cases resolve with fresh drivers; if crashes persist after driver updates and your GPU stays cool, hardware failure is likely.
Key Takeaways
- VIDEO_TDR_FAILURE means your GPU stopped responding within the timeout window, triggering a forced system crash
- Outdated or corrupted AMD drivers are the most common cause, fixable with a clean driver reinstall
- Power management settings can interrupt GPU operations, disable PCI Express Link State Power Management
- Overheating reduces component reliability, clean your GPU and monitor temps below 85°C under load
- If the problem persists after software fixes, hardware failure is likely and GPU replacement may be necessary
At a Glance
- Difficulty: Intermediate
- Time Required: 30-45 minutes
- Success Rate: 78% resolve with driver update alone
What Causes VIDEO_TDR_FAILURE atikmpag.sys Crashes?
Right, let's get technical for a moment. The atikmpag.sys file is AMD's kernel-mode graphics driver. It's the bridge between Windows and your GPU, handling every bit of graphics processing. When Windows boots up, it expects this driver to respond within a tight timeout window, typically around 2 seconds. If it doesn't answer, Windows assumes the hardware is hung and forces a system crash to prevent your computer from locking up permanently. That's the VIDEO_TDR_FAILURE error you're seeing. TDR stands for Timeout Detection and Recovery, and it's a safety net that works by crashing your system rather than letting it freeze.
There are four main reasons this happens. First, and most common, is outdated or corrupted driver software. Windows updates, BIOS changes, or simple file corruption can break the atikmpag.sys file or make it incompatible with your current system. Second is overheating. Dust clogs your GPU fans and heatsink, temperatures climb, and the GPU becomes unreliable under load, exceeding the TDR timeout. Third is power management. Windows might force your GPU into a low-power state inappropriately, breaking communication between driver and hardware. Fourth, least common, is actual hardware failure, a faulty GPU or degraded video memory chips that cause intermittent timeout failures.
The reason crashes often happen during gaming or video playback is simple: that's when your GPU works hardest. Gaming and streaming put heavy load on the card, increasing power draw, heat, and driver activity. If there's a weak spot in your drivers, cooling, or power delivery, intensive tasks expose it. A system might run fine at idle but crash the moment you load a game, which is why the problem feels sporadic and frustrating.
Quick Fix: Update AMD Graphics Drivers
This one fixes the majority of VIDEO_TDR_FAILURE cases, so start here. If your system is stable enough to boot normally, this takes about 20 minutes. If you're stuck in a boot loop, you'll need to boot into Safe Mode first, but the process is identical.
Update AMD Drivers Cleanly Easy
- Boot into Safe Mode if you're experiencing repeated crashes
Restart your computer and press F8 repeatedly during boot. On Windows 10/11, you can also hold Shift while restarting and go to Troubleshoot > Advanced options > Startup Settings, then select Safe Mode with Networking. Safe Mode loads only essential Windows drivers, bypassing the problematic AMD driver entirely. - Uninstall the current AMD driver
Right-click your Start button, select Device Manager. Expand the Display adapters section. Right-click your AMD graphics card and select Uninstall device. Tick the box that says Delete the driver software for this device if it appears. Restart your computer. Windows will now use its basic built-in display driver. - Download fresh AMD drivers from AMD.com
Go to AMD's official support page. Enter your GPU model (you can find this in Device Manager by right-clicking the display adapter and checking Properties, or by using GPU-Z). Download the latest driver package for your exact Windows version. Do not use Windows Update for this; Windows Update drivers are often outdated. - Install with a clean install option
Run the downloaded installer as administrator. Look for an option that says Clean Install or Factory Reset. If it appears, select it. This removes any leftover driver files before installing fresh. Follow the installation wizard and restart when prompted. - Verify stability
After restart, open Device Manager again and confirm your AMD GPU is listed with the correct driver. Now stress test: play a video for 5 minutes, or if you game, load up a game for 15 minutes. Watch for any BSOD. If the system stays stable, you're done.
More VIDEO_TDR_FAILURE Solutions
If driver updates didn't solve it, the next culprits are power management settings and specific driver file corruption. Power management can be surprisingly disruptive. Windows power-saving features sometimes force your GPU into a low-power state while it's trying to process graphics, which causes the driver to timeout. Also, sometimes a single driver file (atikmpag.sys itself) becomes corrupted while others are fine, so replacing just that file can work when a full reinstall doesn't.
Disable Power Management and Replace Driver File Intermediate
- Disable PCI Express Link State Power Management
Open Control Panel. Search for Power Options. Click Change plan settings next to your current power plan. Click Change advanced power settings. Scroll down and expand PCI Express. You'll see Link State Power Management. Set both On battery and Plugged in to Off. Click Apply, then OK. This prevents Windows from cutting power to your GPU during graphics tasks. - Back up your current atikmpag.sys file
Open File Explorer and navigate to C:\Windows\System32\Drivers. Find atikmpag.sys. Right-click it and rename it to atikmpag.sys.old. This preserves the original in case something goes wrong. - Extract the new atikmpag.sys from your AMD driver package
Locate the compressed AMD driver file you downloaded earlier (usually named something like Adrenalin-XX.XX.XX.exe or similar). Inside that package is a compressed file called atikmpag.sy_ (note the underscore at the end). Open Command Prompt as administrator. Navigate to your Downloads folder by typing cd Downloads and pressing Enter. Run this command: expand.exe atikmpag.sy_ atikmpag.sys (adjust the path if needed). This decompresses the file. - Copy the new file to System32\Drivers
Copy the newly extracted atikmpag.sys from your Downloads folder to C:\Windows\System32\Drivers. You may be prompted for administrator permission or asked to confirm replacement. If Windows won't let you copy due to file permissions, restart into Safe Mode and try again. - Restart and test
Restart your computer normally. Run a GPU stress test (play a video, load a game, or use a benchmark tool like FurMark). Monitor the system for 15 minutes. If no crashes occur, the file replacement worked.
Advanced VIDEO_TDR_FAILURE Fixes
At this point, if driver updates and power management tweaks haven't worked, the problem is either hardware-related or thermal. Before you assume your GPU is dead, let's rule out overheating. A surprising number of VIDEO_TDR_FAILURE cases I see through remote support are simply dust. Dust blocks airflow, temperatures climb above 85°C under load, and the GPU becomes unreliable. Once you clean it and cool it down, the crashes stop. Hardware failure is less common but possible, especially in older cards or systems that have run hot for years.
Clean GPU Hardware and Manage Thermals Advanced
- Power down and prepare safely
Shut down your computer completely. Unplug the power cable from the back of the case. Press the power button for 10 seconds to discharge any residual power. Put on an anti-static wrist strap (you can buy these for a few quid online), clip it to a metal part of the case, or touch the metal case regularly to ground yourself. Open your case using the appropriate screwdriver. Avoid touching circuit boards, and work slowly to avoid accidents. - Clean the GPU fans and heatsink
Locate your graphics card and use compressed air (from a can or electric duster) to remove dust. Hold the fans stationary with your other hand while cleaning so the fans don't spin freely, which can damage the bearings. Use short bursts from about 10-15cm away. Pay special attention to the heatsink fins, where dust accumulates thickest. Blow from the inside of the case toward the outside. - Clean additional cooling components
While you're in there, clean the CPU heatsink, case fans, and RAM modules the same way. Remove dust from air intake and exhaust vents. This improves overall case airflow, which keeps your GPU cooler even at idle. - Check GPU seating and power connectors
Look at where your GPU plugs into the motherboard (the PCIe slot). There should be no visible gaps. If the card looks loose, release the retaining clip at the end of the PCIe slot, remove the card, and reseat it firmly. Check that all power connectors (usually 6-pin or 8-pin) are fully inserted and secure. A loose connection causes intermittent power failures that can trigger TDR timeouts. - Reassemble and monitor temperatures
Close the case and reconnect the power cable. Boot up normally. Download and install HWMonitor or GPU-Z to monitor GPU temperature in real time. Run a stress test: load a game or use FurMark for 15 minutes. Your GPU should idle at 30-50°C and stay below 85°C under load. If temperatures are still above 85°C after cleaning, the heatsink may need thermal paste replacement (an advanced procedure) or the GPU may be failing.
If cleaning doesn't lower temperatures or stop the crashes, or if temperatures are already normal and crashes persist, hardware failure is likely. At this point, you've ruled out software and thermal issues. The GPU itself may have faulty VRAM, damaged power delivery circuits, or a failing GPU core. You can run Windows Memory Diagnostic to check system RAM, but the GPU itself will need professional diagnostics or replacement.
I should also mention that if you're experiencing these issues with recording software like AMD ReLive, there can be driver conflicts specific to that feature. If you use ReLive, check our AMD ReLive recording guide to see if disabling that feature helps rule out software-specific causes before you assume hardware failure.
Preventing VIDEO_TDR_FAILURE atikmpag.sys in the Future
Once you've fixed this, keeping it fixed is straightforward. Update your AMD drivers monthly. Go to AMD.com directly, not Windows Update, and check for new releases. Windows Update drivers lag behind by weeks or months, and missing updates means missing stability improvements. I do this on the first of every month as a habit.
Monitor your GPU temperatures at least quarterly. Spend 5 minutes running a stress test and watching HWMonitor. If idle temps are creeping up toward 60°C or load temps are above 80°C, it's time to clean your system. Don't wait for a crash to tell you something's wrong.
Clean your PC internals every 3-6 months, more often if you have pets or live in a dusty area. Takes 30 minutes, prevents weeks of instability. Keep your case fans in good working order, make sure your GPU has unobstructed airflow, and avoid power-saving settings that interfere with GPU operation. Use Windows Balanced or High Performance power plans, not Power Saver.
Avoid GPU overclocking unless you're experienced and prepared to do serious stability testing. Every MHz you push increases power draw and heat, which makes timeout failures more likely. And create system restore points before major driver updates. If a new driver causes problems, you can roll back in minutes rather than spending hours troubleshooting.
VIDEO_TDR_FAILURE atikmpag.sys Summary
VIDEO_TDR_FAILURE atikmpag.sys crashes are frustrating, but they're almost always fixable. Start with a clean AMD driver installation from AMD.com, which solves about 75% of cases. If that doesn't work, disable PCI Express power management and consider replacing the atikmpag.sys file directly. If crashes still occur and your GPU runs cool, clean your hardware and monitor temperatures closely. If cleaning doesn't help, you're looking at hardware failure and GPU replacement.
The key takeaway is this: these crashes are a symptom, not a destiny. Work through the fixes methodically, test each one, and you'll find the culprit. Nine times out of ten, it's software. And software is fixable with a bit of patience and the right steps.