The screen goes black. The crash reporter pops up. “Low Level Fatal Error” or “GPU Crash Dump Triggered” stares back at you. It’s the kind of error that makes even seasoned PC gamers sigh. And the first thing everyone does? Update the GPU driver.
Don’t.
That driver update probably won’t fix this. The real fix is almost certainly a corrupted shader cache. And ignoring that order of operations is why most players end up reinstalling the game or buying new hardware they don’t need.
This isn’t a generic “try everything” guide. It’s a targeted, evidence-backed approach that prioritizes the single most effective step first.
Understanding the GPU Crash Dump Error
The “GPU crash dump triggered” message in Star Wars Zero Company is not a random failure. It signals that the graphics driver encountered an unrecoverable condition during shader compilation. The game asks the GPU to build and run a specific set of instructions for lighting, shadows, and effects. When those instructions are corrupted or incomplete, the driver throws a fatal error.
The crash reporter then collects data from the GPU’s crash dump — a snapshot of the last milliseconds before failure. This is often confused with a hardware problem, but in the vast majority of cases, it’s a software issue.
Two common error variants:
- Low-Level Fatal Error – The driver itself crashed.
- GPU Crash Dump Triggered – The driver survived but the game’s rendering thread died.
Both have the same root cause: problematic shader data.
The Main Culprit: Corrupted Shader Cache
Shader cache is a storage folder where the game saves compiled shaders so it doesn’t have to rebuild them every time. That’s a performance win. But if the cache becomes corrupted — due to a crash during compilation, an incomplete update, or even a system restart at the wrong time — the game will keep trying to load the broken data. Every launch triggers the same crash.
The fix is brutally simple: clear the shader cache.
How to do it:
- Windows Disk Cleanup
- Press Windows Key + R, type
cleanmgrcodecodecodecodecodecode, select the drive where the game is installed. - Click “Clean up system files”.
- Check DirectX Shader Cache.
- Click OK to delete. This removes all cached auto-generated shaders across games.
- Manual deletion (alternative)
- Navigate to
%AppData%LocalTempShaderCachecodecodecodecodecodecode and delete everything inside. - Also check
%AppData%LocalNVIDIADXCachecodecodecodecodecodecode or equivalent AMD/Intel folder.
After clearing, restart your PC. The first launch of Star Wars Zero Company will take longer — shaders must be rebuilt from scratch. That’s normal. And when it finishes, the crash is usually gone.
This is the single most effective fix for the GPU crash dump error. It costs nothing. It requires no downloads. And it works even when everything else fails.
Counterargument: “Aren’t Outdated Drivers the Real Cause?”
It’s a reasonable assumption. Many crash guides lead with “update GPU driver.” AMD, Nvidia, and Intel all push new drivers specifically for new game titles. So surely a fresh driver is the cure.
Here’s the problem: updating a driver does not clear the shader cache. The cache persists across driver versions. If the cache is corrupt, a new driver will still load that corrupt data and crash the same way. You’ll think the new driver didn’t help — and you’d be right to think that, but for the wrong reason.
The strongest evidence comes from player reports across Steam forums and technical support threads. In case after case, updating the driver without clearing the cache fails. Clearing the cache without updating the driver succeeds. The driver itself is rarely the root cause; it’s the victim of bad input.
Of course, old drivers can cause other crashes — instabilities, performance drops, missing features. So updating is still a good idea, but it should be step two, not step one. Leading with the driver update wastes time and creates false negatives.
Step-by-Step Fix Prioritized
Do these in order. Don’t skip.
Step 1: Clear Shader Cache (as described above)
Step 2: Verify Game Files
- Steam: Right-click Star Wars Zero Company → Properties → Installed Files → Verify Integrity of Game Files.
- This checks for missing or damaged game assets. Sometimes shader data is stored in the game’s own files, not just the system cache.
Step 3: Update GPU Driver
- Use the official tool for your GPU: GeForce Experience, AMD Adrenalin, or Intel Driver & Support Assistant.
- After install, reboot and launch.
Step 4: Disable Overlays
- Turn off Discord overlay, Nvidia app overlay, Steam overlay, Xbox Game Bar.
- Overlays inject code into the rendering pipeline, which can conflict with shader compilation.
Step 5: Adjust Launch Options
- In Steam, right-click game → Properties → General → Launch Options.
- Add
-dx11codecodecodecodecodecode (forces DirectX 11) or-dx12codecodecodecodecodecode (DirectX 12). One often works better than the other. - Also try
-windowed -nobordercodecodecodecodecodecode to rule out fullscreen issues.
Step 6: Increase Virtual Memory
- Search “advanced system settings” → Performance → Advanced → Virtual Memory.
- Set custom size to 1.5x your RAM. This prevents out-of-memory errors that can cause shader compilation failures.
Step 7: Run System File Check
- Open Command Prompt as Admin, type
sfc /scannowcodecodecodecodecodecode. Then runDISM /Online /Cleanup-Image /RestoreHealthcodecodecodecodecodecode. - Corrupted Windows system files can indirectly crash shader compilation.
FAQ
Q: What exactly is a GPU crash dump triggered error?
A: It means the game’s graphics driver encountered a fatal error during rendering, and the driver has saved a crash dump file for analysis. In Star Wars Zero Company, this is almost always tied to shader compilation failure.
Q: Can I delete shader cache manually?
A: Yes. The main locations are %AppData%LocalTempShaderCachecodecodecodecodecodecode (system-wide) and %AppData%LocalNVIDIADXCachecodecodecodecodecodecode (Nvidia-specific). Delete all contents, then restart.
Q: Will clearing shader cache affect performance in other games?
A: Only temporarily. Other games will rebuild their own cache on next launch, which may cause brief stutter during the first play session. After that, performance returns to normal.
Q: Should I use -dx11 or -dx12 launch option?
A: Try -dx11 first. Star Wars Zero Company uses Unreal Engine 4, which historically runs more stably on DX11 for shader compilation. DX12 offers better frame times but can introduce driver-level crashes in some hardware combinations.
Q: I cleared cache and verified files but still crash. What next?
A: Update GPU driver, disable overlays, and increase virtual memory. If the crash persists, check your system’s power supply and GPU temperatures. A failing PSU can trigger crash dumps during GPU load.
One More Thing: The Edge Case Nobody Mentions
All of the above works for standard configurations. But there’s a specific scenario where it fails: when the game’s shader compilation pipeline is fundamentally incompatible with certain CPU/GPU architectures. For example, early RTX 3000 series cards paired with Ryzen 5000 CPUs show a higher-than-normal incidence of shader compilation crashes in Unreal Engine 4 games. In that case, no amount of cache clearing or driver updating will fully eliminate it.
The reason is a known interaction between the game’s parallel shader builds and the CPU’s thread scheduler. The game sends too many compilation requests at once, overwhelming the driver’s timeout protection. The result: the driver kills the rendering thread and throws “GPU crash dump triggered.”
The only permanent fix for that edge case is a game patch that throttles the shader compilation rate. Until the developer addresses it, players can manually limit CPU threads by adding -LimitThreads=4codecodecodecodecodecode or -NoAsyncLoadingThreadcodecodecodecodecodecode to the launch options. Not every system needs this, but if you’re on the right (wrong) hardware, it’s the difference between a game that crashes every five minutes and one that runs stable.
That’s the nuance most guides miss. It’s not just about cache or drivers. Sometimes the game itself is asking too much, too fast. And only a developer can fix that.
