Skip to content
Affordable Computer Repair logoAffordable Computer Repair

GPU Artifacts and Crashes: Reading the Visual Symptoms

Coloured blocks, flashing pixels, corrupted textures and display-driver crashes point at the graphics path. Here is how to tell driver problems from heat, power and genuine GPU faults.

Written by The Affordable Computer Repair Bench, Repair TechniciansReviewed by Steven W. Nickerson, Sr.Updated 9 min read

Short answer

Visual corruption on screen means something in the graphics path is producing wrong data — but that path includes the driver, the card's memory, its power delivery, the cable and the monitor. Artifacts do not prove the card is dead.

What artifacts look like

Artifacting is a broad word for the picture going wrong in ways that have nothing to do with the game's own graphics. The pattern matters, so describe it precisely rather than saying the screen 'glitched'.

  • Flashing or dancing pixels scattered across the image
  • Solid coloured blocks or squares, often in a repeating grid
  • Horizontal or vertical lines running across the whole display
  • Textures that look scrambled, stretched or replaced with noise
  • Long spikes of geometry shooting off models — often called polygon spikes
  • The screen going black for a second and returning with a notice that the display driver stopped responding and recovered
  • A hard black screen under load with fans spinning up, needing a restart

What can produce it

AreaTypical presentation
Driver corruption or a bad driver versionDriver-reset messages, artifacts that started right after a driver update, problems in some applications and not others
Unstable overclockArtifacts and crashes only under heavy load, clean behaviour at stock clocks, often after someone tuned the card or enabled a factory OC profile
OverheatingPicture is clean for the first few minutes and degrades as the card heats up; fans loud; temperatures near the card's limit
VRAM faultsBlocky, repeating corruption; artifacts worst in memory-heavy games or at high resolutions and textures
Power deliveryBlack screens and hard crashes under sudden load spikes; can involve the card's own power stages or an ageing or undersized power supply
Cable, port or monitorLines, sparkle or dropouts that change when you swap cable or port; corruption present in BIOS as well as Windows
GPU core or solder-level damageCorruption present from the moment the machine posts, sometimes in the BIOS screen, unchanged by drivers or clocks

Checks worth doing before you buy anything

  1. 1

    Swap the cable and the port

    A tired DisplayPort or HDMI cable produces sparkle and dropped lines that look alarming and cost nothing to rule out. Try a different cable and a different output on the card.

  2. 2

    Check whether it appears before Windows loads

    Restart and watch the manufacturer logo and BIOS screen. Corruption there happens without any driver involved, which moves suspicion firmly onto hardware or cabling.

  3. 3

    Return clocks to stock

    Disable any overclock, factory OC profile or undervolt in the tuning software. If the artifacts stop, the card was being pushed past what it is stable at — which is a settings problem, not a dead card.

  4. 4

    Watch temperatures while it happens

    Run monitoring alongside the game. Corruption that appears only once the card reaches its limit is a cooling problem wearing a graphics-fault costume.

  5. 5

    Do a clean driver reinstall

    Fully remove the existing display driver rather than installing over the top, then install the current release for your card from the manufacturer. Layered driver installs are a genuine source of instability.

  6. 6

    Test a second application

    If one game artifacts and everything else is spotless, suspect that title's shader cache or its own bugs before condemning hardware.

  7. 7

    Reseat the card and its power connectors

    With the machine off and unplugged, confirm the card is fully home in the slot and every required PCIe power connector is firmly clicked in — including the second one people forget.

What artifacts do not prove

Plenty of cards that arrive described as 'artifacting and dying' turn out to be stable once the clocks are at stock, the cooler is clean and the driver stack is rebuilt from scratch. Visual corruption tells you the graphics path is producing bad output. It does not tell you the silicon is permanently damaged, and it does not on its own justify buying a replacement card.

  • It does not prove the GPU is dead — drivers, heat, clocks and cables all produce the same symptoms.
  • It does not prove the power supply is fine either; a marginal PSU shows up as crashes under load.
  • It does not rule out the monitor, which is the cheapest thing on the list to test.

When to hand it over

If corruption survives a clean driver rebuild, stock clocks, a known-good cable and healthy temperatures, the next step is testing the card in another machine and testing another card in yours. That is the point most people run out of spare parts, and it is exactly what a bench is for. Bring it to the shop in Bellevue and we will isolate whether the card, the board or the power supply is at fault before anyone spends money on a replacement.

Common questions

Can a driver update really cause artifacts?
Yes. A bad install, a leftover from a previous driver, or a release with a regression for your card can all produce corruption and driver resets. A clean removal and reinstall is one of the first things worth trying.
My card artifacts only in one game. Is it faulty?
Probably not. Single-title corruption usually points at that game's shader cache, its own bugs, or an unstable overclock that only that engine loads hard enough to expose.
Is it safe to keep using a card that artifacts?
It won't harm you, but crashes can corrupt open files and interrupt Windows updates mid-write. Back up anything important before you continue troubleshooting.
Do artifacts mean the VRAM is bad?
VRAM faults are one cause, and they tend to show as blocky, repeating patterns that get worse at high texture settings. That pattern is suggestive, not conclusive — it still needs testing against known-good hardware.

The service that covers this

Gaming PC Repair & Upgrades

Custom builds and gaming rigs — diagnostics, component swaps, cooling, and performance work.

Gaming PC Repair & Upgrades in Bellevue →

Related guides

← More gaming pc repair

Free counter consultation

Call NowBook Repair