How to Tell if a GPU Was Used for Mining

Buying a used graphics card can save you a lot of money, but there’s a chance it spent years mining cryptocurrency. Mining puts constant stress on a GPU’s memory and cooling system, which can shorten its lifespan. This guide shows you how to inspect a GPU for signs of mining use, both physically and through software, so you can make an informed purchase.

What You’ll Need

• The GPU itself (or clear photos if buying online)
• A computer to test the GPU in (or a friend’s rig)
• Free software: GPU-Z, HWiNFO64, and a stress test like FurMark or Unigine Heaven
• A small Phillips-head screwdriver (if you plan to inspect the cooler)
• Isopropyl alcohol and cotton swabs (for cleaning if you buy it)

Step-by-Step

1. Check the seller’s story. Ask directly: “Was this card used for mining?” A honest seller will tell you. If they hesitate or give a vague answer like “light gaming only,” be suspicious. Miners often sell in bulk, so if the seller has multiple identical cards, that’s a huge red flag.
2. Look at the physical condition. Remove the card from its box or anti-static bag. Look for:
   • Discoloration or yellowing on the backplate or PCB (printed circuit board) near the power connectors – heat damage.
   • Dust buildup in the heatsink fins, especially if it’s caked on and hard to remove.
   • Scratches on the PCIe connector or screw holes – signs of multiple installs/removals.
   • Missing or replaced thermal pads (sometimes visible between the PCB and backplate).
   • Any stickers or markings that say “LHR” (Lite Hash Rate) – these were made to limit mining, so if present, the card was likely aimed at gamers, though not a guarantee.
3. Inspect the cooler and fans. Spin the fans by hand. They should spin freely without grinding or wobbling. If you can, remove the shroud (usually 4-6 screws) and check the thermal paste. Fresh, evenly spread paste is good; dried, cracked, or flaky paste suggests heavy use. Also look for corrosion on the heatsink – a sign of humid mining environments.
4. Install the GPU and run GPU-Z. Put the card in a test system, install drivers, and open GPU-Z. Look at the “Sensors” tab. Key things to note:
   • GPU Clock and Memory Clock: At idle, they should drop to low values (e.g., 300 MHz core, 200 MHz memory) if power management is working. If they stay maxed out, the card may have a modified BIOS from mining.
   • GPU Temperature: At idle, it should be near room temperature (30-40°C). If it’s hotter, the cooler may be clogged or the thermal paste degraded.
   • Memory Temperature (if shown): GDDR6X memory on high-end cards runs hot, but at idle it should be under 60°C. If it’s above that, the thermal pads are likely worn.
5. Check the BIOS version. In GPU-Z, click the “BIOS Version” field. A mining BIOS often has a different version number or a date that’s older than the card’s release. If the BIOS version doesn’t match the official one from the manufacturer’s website, it’s been flashed – a common mining practice. You can also click the “Lookup” button to compare.
6. Run a stress test. Download FurMark or Unigine Heaven and run it for 15-20 minutes. Watch the temperatures and clock speeds. A healthy GPU should:
   • Reach its boost clock (within 50 MHz) and hold it without dropping.
   • Stay below 85°C for most cards (check the specific model’s limit).
   • Not crash, artifact (show visual glitches), or produce a black screen.

   If the card throttles hard or overheats quickly, it’s been stressed and may be failing.

7. Check for artifacts and errors. While the stress test runs, look for flickering, colored dots, or lines on the screen. Also, run a memory test like OCCT’s VRAM test or use HWiNFO64 to check for “Memory Errors” or “WHEA Errors” – these indicate failing memory modules, common in mined cards.
8. Listen and feel. Put your hand near the back of the card (where the bracket is) to feel the heat. A card that’s been mined will often have a distinct hot spot near the memory chips. Also, listen for coil whine – a high-pitched squeal under load. While not exclusive to mining, it’s more common in cards that have been run at high loads for long periods.
9. Ask for the mining history. If the seller admits mining, ask for how long and at what settings. A card mined for 6 months with undervolting and good cooling might be fine; one mined 24/7 for 2 years with high temperatures is a risk. Also ask if the card was used with a riser cable – that’s typical in mining rigs.
10. Decide based on evidence. No single sign proves mining, but a combination of red flags (multiple cards, modified BIOS, high idle temps, missing thermal pads, poor stress test results) should make you walk away. If the card passes all tests and looks clean, it’s probably fine, regardless of its past.

Common Mistakes

• Relying only on the seller’s word. Some sellers lie or don’t know the history. Always verify with software.
• Ignoring idle temperatures. A card that’s hot at idle is a clear sign of poor cooling, which mining causes.
• Skipping the stress test. A quick boot and a glance at the desktop isn’t enough. You need to push the card to see if it fails.
• Assuming all mining cards are bad. Some miners undervolt and maintain their cards well. A well-cared-for mining card can be a great deal. The key is to test, not assume.
• Forgetting to check the warranty. Most manufacturers void warranties if the card was used for mining. If the seller says “warranty still active,” verify with the manufacturer using the serial number.

Tips

• Buy from a reputable reseller. Platforms like eBay have buyer protection, and some sellers offer returns. A private Facebook Marketplace deal is riskier.
• Ask for a video of the card running a stress test. If the seller won’t provide one, that’s a red flag.
• Check the card’s serial number online. Some manufacturers have a warranty checker that shows the original purchase date. If the card was bought 3 years ago but the seller says “barely used,” be skeptical.
• Consider the price. If the card is significantly cheaper than market value, there’s likely a reason. Don’t let a low price blind you to potential issues.
• Clean and repaste the card if you buy it. Even if it looks fine, replacing the thermal paste and pads can restore performance and extend life. There are many guides online for your specific model.

Troubleshooting

• The card crashes during the stress test. This could be a power supply issue (not enough wattage), a driver problem, or a failing GPU. Try a different PSU or reinstall drivers with DDU (Display Driver Uninstaller). If it still crashes, the card is likely defective – return it.
• GPU-Z shows a different BIOS version than expected. Search the BIOS version online. If it’s not the official one, you can try flashing the original BIOS (find it on TechPowerUp’s VGA BIOS database). But be careful – a failed flash can brick the card. Only do this if you’re comfortable with the process.
• The card runs fine but temperatures are high. Replace the thermal paste and pads. This is a standard maintenance task and often fixes overheating issues.
• You see artifacts but the card passes the stress test. Run a more intensive test like OCCT’s VRAM test. If it fails, the memory chips are damaged – this is a common mining failure. Do not buy the card.
• The seller refuses to let you test the card. Walk away. A legitimate seller will allow testing, especially if you’re meeting in person. If buying online, ask for a video of the card running a stress test with the CPU-Z and GPU-Z visible.

Frequently Asked Questions

Can a GPU be mined on without any physical signs?

Yes, if the miner used good cooling and undervolting, the card might look clean. That’s why software checks are essential – a mining BIOS or memory errors won’t show on the outside.

Is it safe to buy a used mining GPU?

It can be, if the card passes all tests and the price is right. However, there’s a higher risk of failure than with a gaming card. Many mining cards are sold with no warranty, so you’re taking a gamble.

What does LHR mean and does it matter?

LHR stands for Lite Hash Rate, a limiter Nvidia added to GPUs to discourage mining. LHR cards are less likely to have been mined on, but miners found ways to bypass the limiter, so it’s not a guarantee.

How long do GPUs last if used for mining?

It depends on the conditions. A card mined 24/7 at high temperatures can fail within a year, while a well-cooled card might last several years. The memory chips are usually the first to fail.

Can I fix a GPU that was used for mining?

Sometimes. Replacing thermal paste and pads can fix overheating. If the memory is failing, it’s usually not worth repairing – the cost can be close to a new card.

Should I avoid all used GPUs, just in case?

No, many used GPUs are from gamers and are perfectly fine. Just do your due diligence – ask questions, inspect, and test. A used card can offer great value if you’re careful.

Related

• Wave Browser: What It Is and How It Works
• How to Block Chrome: Stop It From Opening
• Solved Windows 10 Freezes Locks Up Randomly Easily?