A failing graphics card is one of those problems that often gets blamed on something else — a buggy game, a driver update gone wrong, or a tired old PC that just needs a clean install. By the time most people accept the GPU itself is the issue, they've already wasted a weekend reinstalling Windows. If you're seeing graphical glitches, sudden crashes during games, or a black screen on boot, your graphics card is one of the first things to investigate.
Here are the most common warning signs of a failing GPU, the quick tests you can run at home, and what to do next if your card really is on the way out.
1. Visual Artifacts on Screen
Artifacts are the clearest sign of GPU trouble. These are graphical glitches that have no business being there: random coloured dots, horizontal or vertical lines, blocks of garbled texture, flickering shapes, or distorted text. They tend to show up first inside games or 3D apps, but in a badly failing card they'll appear on the Windows desktop too.
If you only see artifacts in one specific game, it's probably a software bug. If you see them across multiple games, on YouTube videos, or even on the Windows login screen, the card itself is the suspect.
2. Frequent Crashes During Games or Video Editing
GPUs are stressed hardest when rendering 3D graphics or transcoding video. A weak or failing card will often throw the system back to desktop, lock up entirely, or trigger a blue screen of death mid-session. If your PC is rock-solid in the browser and Office, but consistently crashes ten minutes into a game in Leith or a 4K render in Morningside, the GPU is a strong candidate.
3. "Display Driver Stopped Responding and Has Recovered"
That little notification in the bottom-right of Windows is the operating system telling you the graphics driver crashed and had to be restarted. The screen usually goes black for a second, then recovers. One-off events happen — but if you're seeing this several times a week, you have a hardware or driver problem that needs sorting.
Always update to the latest stable driver from NVIDIA or AMD before assuming the card itself is faulty. A clean driver reinstall using DDU (Display Driver Uninstaller) in Safe Mode resolves a surprising number of these cases.
4. Loud or Constantly Running Fans
Modern graphics cards have temperature-controlled fans that idle quietly when the GPU is cool. If your card's fans are roaring at full speed even when you're just on the desktop, the card is either overheating or its thermal sensors are misreporting. Both point at the GPU.
Dust build-up is the most common cause. After a couple of years in the typical Edinburgh tenement, a tower picks up enough dust to choke the cooler. A careful clean with compressed air often fixes it. Our PC overheating guide walks through the process in detail.
5. Black Screen on Boot or No Display Output
If your PC powers on, fans spin, but the monitor stays black with a "no signal" message, the graphics card has either died or come loose. Before panicking, check the obvious things first: try a different cable, a different monitor input, and reseat the card in its PCIe slot. If you have integrated graphics on the motherboard, plug the monitor into that and see if Windows boots — if it does, your dedicated GPU is the problem.
6. Drops in Frame Rate or Performance
If a game that used to run at 90 fps now stutters along at 30, and nothing has changed in your settings, the card may be thermal-throttling because of dust, dried-out thermal paste, or a partial fault. Tools like MSI Afterburner can show you the GPU's clock speed in real time — if it's clocking down dramatically under load, heat is the issue.
How to Test Your Graphics Card
Before booking a repair, run a few quick checks:
- Update drivers cleanly. Uninstall the current driver in Safe Mode using DDU, then install the newest one from the manufacturer's website.
- Monitor temperatures. Use HWMonitor or the manufacturer's utility. Idle should sit under 50 °C; under load, most cards stay below 85 °C. Anything significantly higher needs attention.
- Run a stress test. FurMark, 3DMark, or Unigine Heaven will load the GPU hard for 10–15 minutes. If you see artifacts, crashes, or an automatic shutdown, the card has a real fault.
- Swap the card. If you have a spare GPU or a friend who'll lend one, dropping it into your PC immediately confirms whether the original card is the issue.
Repair, Reseat, or Replace?
Some GPU issues are easy fixes — dust, a loose card, an outdated driver, or a tired thermal pad. Others, like dead VRAM chips or a damaged GPU die, need component-level work. Microsoldering and board-level repair can sometimes save a high-end card; in other cases, replacement is the cheaper route. Our hardware upgrades service covers GPU swaps, and if you're rebuilding for gaming, take a look at our custom PC builds for matched configurations.
How We Can Help in Edinburgh
From Portobello to Dalkeith, we diagnose graphics card problems every week. We'll bench-test your card, rule out drivers and PSU issues, and tell you honestly whether it's worth repairing or time to upgrade. Whether you bring the PC to our workshop or use our home callout service, you'll get a clear answer fast. Book a diagnostic online or message us on WhatsApp.