Quick answer: For blocks, lines, or flicker, compare a known-good compatible display cable or monitor and restore any unstable overclock. If artifacts persist across sources or the card overheats, stop and seek qualified diagnosis.
Before you start
- Tools: the product manual, good lighting, a soft dry cloth, and one known-good compatible cable or accessory if relevant
- Time: 15–45 minutes for external checks
- Difficulty: Moderate
- Safety: Shut down, switch off and unplug the PC before reseating a card. Use ESD precautions. Do not open a power supply.
Back up accessible data before updates. Work on a stable, dry surface and disconnect power before moving cables or removable hardware.
Causes ranked by likelihood
| Cause | How to check | Fix | Difficulty |
|---|---|---|---|
| 1. Cable or monitor fault | Compare with a known-good compatible cable or screen | Replace a damaged external cable | Easy |
| 2. Unstable overclock or memory | Note changes after an optional performance profile | Restore documented defaults and retest | Easy |
| 3. Driver or application issue | Note if artifacts appear in one app or systemwide | Use a driver matched to the exact GPU and OS | Moderate |
| 4. Failing GPU memory or hardware | Note persistent artifacts across sources | Stop stressing the system and seek diagnosis | Pro |
Key takeaway: 1. Cable or monitor fault: how to check — Compare with a known-good compatible cable or screen.
1. Check cable or monitor fault
Shut down both ends, inspect for bent contacts or strain, and reconnect the correct port firmly. Test one known-good cable that supports the required signal or power standard. Adapters and hubs add another failure point, so remove them for the isolation test.
2. Isolate unstable overclock or memory
If artifacts began after GPU or memory overclocking, restore those settings to the hardware maker's defaults. Compare a normal workload without stressing a card that shows worsening artifacts or electrical odor.
3. Evaluate driver or application issue
Record the current version, then obtain the model-correct package from the hardware or operating-system vendor. Restart after installation. If the fault began with an update, use the vendor's documented rollback; repeated reinstalling will not correct unstable hardware.
4. Treat failing gpu memory or hardware as an escalation point
Persistent artifacts across multiple displays or clean driver installations may indicate a GPU hardware fault. Stop intensive testing and seek qualified diagnosis; do not open the PSU or attempt board-level repair.
Brand-specific notes
Driver packages and supported outputs vary by GPU and system. Use the exact manufacturer's support page to identify a compatible driver.
When to call a professional
Arrange qualified service for a burning smell, smoke, arcing, liquid inside, damaged mains connector, swollen battery, repeated protection shutdown, precision alignment, soldered port, or a fault that requires energized measurements. For valuable data on a failing drive or card, stop writing to it and consult a recovery specialist.
Prevention tips
- Keep vents, optics, contacts, and connector openings clean using the maker's approved method.
- Avoid cable strain, impact, moisture, extreme temperatures, and unsupported chargers or adapters.
- Maintain backups before faults occur and record stable settings before updates.
- Inspect lithium batteries regularly; retire damaged or swollen packs without charging or reuse.
Related help
- Back to Graphics Card Guide
- GPU Not Detected? Safe Checks That Find the Cause
- GPU Overheating: Temperatures, Airflow, and Fixes
- Electronics safety basics
Official resources
Summary
Begin with the most likely external cause, verify the complete power or signal path, and test one known-good compatible item at a time. Model-specific documentation outranks generic reset advice. Stop before a low-risk check becomes internal powered work, battery damage, precision mechanical damage, or data loss.
Observations that narrow down gpu artifacts
| What you observe | What it may indicate | Safe next step |
|---|---|---|
| Pattern appears only on one display or cable | Signal path may be responsible | Try one known-compatible cable or input with devices powered down. |
| Pattern appears only in one application | Software settings may be relevant | Return overclocks to supported defaults and compare another application. |
| Artifacts occur before the operating system loads | Hardware becomes more plausible | Record a photo and seek qualified diagnosis. |
Key takeaway: Pattern appears only on one display or cable: what it may indicate — Signal path may be responsible.
Key takeaway: For gpu artifacts, compare the observation with one safe, reversible check before assuming a part has failed.