How to Check PC Temperature While Gaming: CPU and GPU Guide

Checking PC temperature while gaming takes about two minutes once a sensor overlay is on screen. Read the CPU package temperature and the GPU core and hotspot temperatures after roughly 15 minutes of steady play, then compare them with the maximum listed for your exact chips. For most gaming rigs, a sustained CPU reading below about 85C and a GPU core below about 80C to 85C leaves a comfortable margin, although some cards run hotspots several degrees higher and still behave normally. Treat these figures as rough estimates, because every model has its own limit.

Owen Pritchard, our Technical Guides Editor with 12 years of repair work behind him, checks these readings on a diagnostics bench with POST cards, spare parts for swap-testing and multiple OS images. The ranges in this guide are estimates drawn from typical manufacturer limits. They are not numbers recorded from a controlled run, so the datasheet for your own CPU and GPU is the final reference.

What each temperature reading actually means

A modern CPU reports several sensors, but the one that matters most in a game is the package temperature, which reflects heat across all cores together. Individual core readings can spike higher for a moment when one thread is busy, and that is normal. If you only watch one number, watch the package value. On Intel and AMD chips it is usually labeled “CPU Package” or “Tctl/Tdie” in monitoring tools, and the label is the first thing to confirm before you trust a reading.

GPU readings split into three values. The core temperature is the general die reading. The hotspot, sometimes called junction, is measured at the hottest point of the die and usually runs noticeably above the core value under load. Memory junction temperature matters on GDDR6 and GDDR6X cards, where the memory can run well above the core after long sessions. Knowing which sensor you are looking at prevents most of the false alarms people report when they first start monitoring.

Temperature alone does not show whether the system is slowing itself down. When a chip reaches its thermal limit, it lowers clock speed to protect itself. A temperature that looks acceptable on paper can still come with falling clocks, stutter, and frame-time spikes. The clock column matters as much as the heat column, so log both at the same time.

Free tools that read sensors while you play

HWiNFO is the most complete free option for Windows. It reads CPU, GPU, motherboard, and storage sensors, and its sensor window can sit beside a game. Its logging feature writes a CSV file, which is useful when you want to compare two sessions after a cooler change or a fan curve edit. The interface is dense, so open only the rows you need: CPU Package, CPU clock, GPU Temperature, GPU Hot Spot, GPU Clock, and GPU Memory Junction if your card reports it.

MSI Afterburner, paired with the RivaTuner Statistics Server that installs alongside it, produces an on-screen overlay. You choose which lines appear in a corner of the screen, so a temperature and a clock speed can sit next to your frame rate without a second window. Afterburner works with both NVIDIA and AMD cards. Some CPU sensors may not appear there unless HWiNFO is also running, which is why many people run both.

GPU driver suites from NVIDIA and AMD show core temperature and usage inside their own performance overlays. These are convenient, but they usually expose less detail than HWiNFO and may not show hotspot or memory junction values. Windows Task Manager shows CPU and GPU utilization, but it does not display temperature, so use it to confirm which processes are consuming resources rather than to judge heat.

Motherboard utilities from ASUS, MSI, Gigabyte, and ASRock also report CPU and system temperatures. Their sensor names vary, and some boards report a socket or voltage regulator reading instead of the die value. When a board utility and HWiNFO disagree by a few degrees, trust HWiNFO for CPU package and use the board utility for case and regulator context.

Setting up an in-game overlay step by step

Start by installing HWiNFO and launching it in sensors-only mode. Keep it closed when you do not need it, or set it to start with Windows if you want it always available. Next, install MSI Afterburner, open its settings, and go to the Monitoring tab. Select the values you want, such as CPU temperature, CPU usage, GPU temperature, GPU usage, and frame rate, and tick the option that shows each value in the on-screen display. Place the overlay where it does not cover the crosshair, minimap, or ammo counter in the game you play most.

Confirm the overlay works inside the game before a long session. Most titles need the RivaTuner server running in the background, and a few launchers or anti-cheat systems block overlays altogether. Load a training area or a short level, then check that the numbers refresh about once per second. If they freeze, close and reopen Afterburner with the game shut down, then start the game again. If the overlay still does not appear, log the values with HWiNFO and review the file afterward.

Decide in advance what to watch. A good first set is CPU package temperature, CPU clock speed, GPU core temperature, GPU hotspot, GPU clock speed, fan percentage, and frame rate. Frame rate belongs on the list because throttling often shows up as frame-time instability before it shows up as a temperature problem. Keeping all of these on one screen lets you see cause and effect together, which is the point of the exercise.

How to run a fair 15-minute test session

A temperature reading is only useful when the conditions are repeatable. Start with the PC at room temperature, ideally after it has been idle for at least 20 minutes. Close browsers, launchers, and sync tools, and note the room temperature if you have a thermometer. A room that is three or four degrees warmer can raise component temperatures enough to make two sessions look different even when the hardware has not changed at all.

Choose one demanding but repeatable scene. An open-world area with heavy foliage, a dense city, or a fixed route through a crowded map works better than a menu or a loading screen. Play for 15 minutes without pausing, because the first few minutes show only the early heat rise. Once the load stabilizes, the last five minutes of the session are the numbers to record. Pausing mid-test lets the cooler catch up and hides the real plateau.

Record the maximum and average value for each sensor, not only the peak. HWiNFO can export this data as a CSV file that opens in any spreadsheet. A single high spike during a loading screen matters far less than a sustained average. Write down the fan curve or fan speed setting, whether the case door is open or closed, and whether the power plan was set to balanced or high performance. These notes make it much easier to repeat the test after a change, and they stop you from blaming the wrong variable.

Reading the results: safe, warm, and worth fixing

The table below gives rough reference ranges for sustained gaming loads. These are estimates based on typical vendor limits and common desktop behavior, not measurements from a controlled run, and your exact chip may behave differently.

A reading in the middle row is not automatically a failure. The problem is usually a combination of high temperature, lower clock speeds, and frame-time spikes. If clocks hold steady across the whole session, the cooler is doing its job even when the number looks high. If clocks drop repeatedly while temperature sits near the top of the range, the cooler, the fans, or the airflow needs attention.

Look at the trend rather than only the final value. A temperature that rises steadily for 15 minutes and never levels off usually points to insufficient cooling capacity or a fan that is not spinning up. A temperature that plateaus quickly suggests a stable system, even if the plateau is warm. Compare fan percentage with temperature to see whether the fan curve is too conservative, because many GPUs keep fans quiet until a fairly high temperature.

Common causes of high temperatures during gaming

Dust is the most common cause. A filter or heatsink with a thick layer of lint reduces airflow, and the fans have to work harder to move the same volume of air. A normal home setup usually needs the front intake filter cleaned every one to three months, and more often if you have pets or carpet nearby. Compressed air works well on heatsinks, but hold the fan blades still while blowing so they do not spin past their rated speed. Choosing a filter that is easy to remove makes this a two-minute job. Our guide to choosing the right dust filters covers the options.

Thermal paste that has dried out or was applied poorly is the second most common cause, especially on CPUs several years old. A cooler that was removed and reinstalled without fresh paste can leave an uneven contact patch, which raises package temperature under load. Mounting pressure matters too. If the cooler is not seated evenly, one corner of the die runs hotter than the rest and the sensor reports a higher number than the average.

Airflow design is the third cause. A case with a closed front panel, a top-mounted radiator pulling warm air up from the GPU, or a rear exhaust fan that runs slower than the intake can trap heat. Check whether the fans pull air in from the front and push it out the back and top. The GPU fan curve is another frequent culprit, because many cards keep fans quiet until the core is already warm. A slightly more aggressive curve often lowers peak temperature noticeably without a large increase in noise.

Room temperature and background software round out the list. A PC placed in a warm room or beside a heater starts each session from a higher baseline. Launchers, screen recorders, and browser tabs playing video add load you did not plan for. Close them before a test so the results reflect the game rather than the background workload.

What to do if temperatures keep climbing

Start with the simplest checks. Confirm that every case fan and the CPU cooler fan spin up as the load rises, and that none of them stop, wobble, or rattle. Open the case, clean the filters and heatsinks, and make sure no cable touches a fan blade. If the room is warm, move the PC to a cooler spot or cap the frame rate for one session to confirm whether load or ambient heat is the driver.

If the fans respond but temperatures still climb, remove the cooler and inspect the contact patch on the CPU. A thin, even spread with clean edges usually means good coverage. Dry, cracked, or pooled paste points to reapplication. Follow the mounting order in your cooler’s manual and tighten the screws in a cross pattern to a firm but not excessive torque. Our walkthrough on applying thermal paste covers the process step by step.

If the system still throttles, check the power limits in the BIOS or the vendor utility. Some motherboards raise power limits aggressively, which can overwhelm a modest cooler. Lowering the processor power limit by a small amount often keeps clocks steadier with little loss in frame rate. Update the BIOS if the vendor has released a fix for thermal behavior, and reset to defaults if you have changed several settings at once. For the wider list of symptoms and causes, the overheating fix guide walks through each one.

Stop the session and shut down if the PC restarts during a game, the screen goes black while the fans keep running, or the GPU shows visual artifacts. These can signal a protection shutdown or a power fault rather than simple heat. Leave the PC off until the cause is known, and inspect the power supply and its cables before the next attempt. A failing power supply can mimic heat problems, so compare your symptoms with the power supply diagnosis guide before replacing any cooling parts.

Frequently asked questions

Is 85C too hot for a CPU while gaming?

Not automatically. Many desktop CPUs are designed to run near 90C to 100C before throttling, so a sustained 85C is warm but often acceptable if clock speeds stay steady. Check the maximum temperature listed in your chip’s spec sheet or the vendor’s support page. Real trouble starts when temperature keeps climbing while clocks drop, fans get very loud, or the PC shuts down mid-session.

Which free program is best for checking PC temperature while gaming?

HWiNFO is the most complete free option on Windows because it reads CPU, GPU, motherboard, and storage sensors and can log them to a CSV file. For an on-screen display, MSI Afterburner with its RivaTuner overlay shows temperature, clock speed, and frame rate in the corner of the game. Using both gives you a detailed log and a live readout.

Does Windows 11 Task Manager show CPU temperature?

No. Task Manager on Windows 11 shows CPU and GPU usage, memory, and running processes, but it does not display temperature. To read temperatures you need a sensor tool such as HWiNFO, a motherboard utility, or the overlay in your GPU driver suite. Task Manager remains useful for finding which background process is consuming resources during a session.

How long should I game before recording temperatures?

Play for about 15 minutes of steady, demanding gameplay before recording final values. Temperatures rise quickly in the first few minutes and usually level off after 10 to 15 minutes of consistent load. Record the average and maximum for the last five minutes, and keep room temperature, fan settings, and the game scene the same between sessions so the comparison is fair.

Why is my GPU hotspot much higher than the core temperature?

The hotspot, or junction, sensor measures the hottest point on the GPU die, while the core sensor reports a broader average across the die. Under heavy load the gap between them is normal and can reach several degrees or more. Worry if the hotspot climbs far above the card’s rated range, if the gap keeps widening over repeated sessions, or if the card starts to throttle.