How to Overclock a GPU Safely With MSI Afterburner: A Step-by-Step Walkthrough

To overclock a GPU safely with MSI Afterburner, raise the core and memory clocks in small steps, test stability after each change, keep temperatures in a sensible range, and leave voltage alone unless you know what you are doing. A typical result is a few extra percent of performance, not a transformation, so the goal is a stable gain rather than a record. This guide, written by our editorial team from published specifications, manufacturer documentation and general knowledge, covers preparation, the step-by-step process, signs of trouble and a worked example with stated assumptions. Always check the warranty policy of your card maker first.

Is overclocking worth it?

Many current graphics cards already boost automatically to the highest clock their power and temperature allow, which leaves limited headroom compared with older generations. A manual overclock may still gain a small amount, and the same tool lets you do the opposite: lower power draw and temperature with little speed loss through a gentle undervolt. Think of the process as tuning, with benefits that range from a few extra frames to a cooler, quieter card.

It is not a fix for a graphics card that is too slow for your games. If frame rates are far from your target, lowering settings or upgrading will matter more. Overclocking suits people who already have solid cooling, a healthy power supply and some curiosity.

Prerequisites before you start

Preparation prevents most problems. Work through this list first.

• Update the graphics driver from the chip maker, and make sure Windows is stable.
• Check cooling: clean dust from the case and card, confirm every fan spins, and make sure the case has airflow. Overclocking on a hot card gains little.
• Check the power supply: confirm it meets the card maker’s recommended wattage and has the correct power connectors.
• Download Afterburner from the official MSI website only. It installs alongside RivaTuner Statistics Server for on-screen monitoring.
• Pick a benchmark and a game you can repeat, so you can compare results fairly.
• Record a baseline: note the benchmark score, the maximum temperature and the fan noise at default settings.

Laptops are a different story. Cooling and power are limited, and some manufacturers restrict overclocking, so be especially cautious and check the guidance for your model. Remember that overclocking can affect your warranty, depending on the manufacturer.

Step by step in MSI Afterburner

1. Open Afterburner and confirm it shows your GPU, the current clocks and the temperature graph.
2. Raise the power limit and temperature limit sliders modestly, if the card permits it. This lets the card keep boosting rather than stopping early.
3. Leave the core voltage slider where it is.
4. Increase the core clock by a small amount, such as 25 MHz, and click apply.
5. Run your benchmark or a demanding game for several minutes, watching for crashes, flickering or artifacts.
6. If stable, raise the core clock again by the same small step and repeat. Stop when you see instability, then back off one or two steps.
7. Repeat the process for the memory clock separately, using slightly larger steps if the card tolerates them, and again checking for artifacts or lower scores.
8. Set a fan curve that ramps up sensibly with temperature, balancing noise and cooling.
9. Save the result to a profile and enable the option to apply it at startup only once you are confident it is stable.

The golden rule is to change only one thing at a time. If you adjust the core, the memory and the fans at once, you will not know which change caused a crash.

Testing for stability

A setting that survives five minutes may fail in a long session. Use a repeatable stress test or benchmark loop, then play several different games for an hour or more. Watch the temperature graph and the clock speed. If the clocks fall as the card heats up, the cooling is limiting you more than the settings.

Look for visual artifacts such as flickering textures, odd colors or sparkles, which usually point to unstable memory. Crashes to the desktop, driver resets and black screens often point to an unstable core or insufficient power. A higher benchmark score is not proof of stability, because some cards produce lower scores once memory error correction kicks in. Use the score as a guide and the games as the real test.

Keeping heat and noise under control

More speed means more heat. A card that runs near its thermal limit will lower its own clocks, undoing your work. If temperatures climb, back off the overclock or improve airflow. Another way to cut heat is to limit the frame rate, which keeps the GPU from rendering more frames than the display can show, as explained in our guide to capping FPS to cut GPU temperature and fan noise.

Fan noise matters for daily use. A custom curve that holds the fans at a steady, tolerable speed may be preferable to a curve that jumps up and down. If the card becomes too loud for a small gain, consider a mild undervolt instead.

Worked example with stated assumptions

Suppose a desktop graphics card scores 100 points in your chosen benchmark at default settings and peaks at 70 degrees Celsius. Assume the case airflow is good and the power supply is well above the recommended wattage. You raise the power limit slightly, add 25 MHz to the core and test, then repeat until a few steps in the card shows a flicker. You back down one step, add a modest amount to the memory in the same way, and stop when the benchmark score stops improving.

In a scenario like this, a final result of 104 to 106 points with a temperature increase of a few degrees would be a typical, believable outcome. The numbers here are illustrations, not promises. Your card could gain more or less, and the best result is the one that stays stable for hours.

Troubleshooting and when to stop

If the screen goes black or the PC restarts, reboot, open Afterburner and press reset. Do not set the profile to apply at startup until the settings are proven. If problems persist at default settings, check driver installation, cable seating and power connectors, and see our guide on fixing a GPU that is not being detected if the card disappears.

Stop immediately if you smell burning, see smoke, hear coil whine that worsens, or find temperatures pinned at the card’s limit. Turn off the PC and unplug it before inspecting hardware. If you are not comfortable continuing, return to default settings. The potential gain is small and the downside of hardware damage is not worth it.

Related guides

• Which Motherboards Support Rtx 4090 And Rtx 5090
• How To Lower Cpu Temperature On A Gaming Laptop
• Affordable Laptop For Small Business

Frequently asked questions

Is overclocking a GPU with MSI Afterburner safe?

It is generally considered low risk when done in small steps within the limits the software allows, because modern GPUs throttle and protect themselves. The real downsides are crashes, visual artifacts and extra heat and noise. Avoid raising voltage, keep temperatures in check, and revert to defaults if anything seems unstable. Check your card maker’s warranty terms, since policies on overclocking differ.

How much performance does a GPU overclock add?

Usually a modest amount, often a few percent and sometimes a bit more depending on the card, cooling and silicon quality. Every chip is different, so results vary and no figure is guaranteed. Many current cards already boost close to their limit on their own, which leaves less headroom than older generations. Measure with the same benchmark before and after to see what you actually gained.

What is the difference between core clock and memory clock?

The core clock sets the speed of the graphics processing units, while the memory clock sets how fast the video memory runs. Raising either can improve frame rates, but they fail in different ways. Core instability often causes crashes, while memory instability can cause artifacts or reduced performance from error correction. Adjust one at a time so you can tell which change caused a problem.

Should the voltage slider be raised?

Beginners are usually better off leaving it alone. Raising voltage increases heat and power draw and can shorten the life of components if pushed too far, and many cards limit it anyway. A safer path is to raise the power limit slightly, increase the clocks in small steps, and consider a gentle undervolt for lower temperatures. Read the manufacturer guidance for your card first.

What should be done if the system crashes after an overclock?

Restart, open Afterburner and click reset to return to default settings, and do not apply the profile at startup until the settings are stable. Then reapply a smaller change and test again. If crashes continue at stock settings, the cause may be drivers, power supply capacity, temperature or hardware, and you may need to troubleshoot or contact support.