To monitor CPU and GPU temperatures, install HWiNFO for full sensor detail or MSI Afterburner with the RivaTuner Statistics Server overlay to see readings inside a game, then watch the CPU package and GPU core values against safe ranges. Both chips report temperature from sensors built into the silicon, and a monitoring tool reads those sensors in real time. A safe target is 30 to 45 degrees Celsius idle and 65 to 85 degrees Celsius under load for the CPU, and 30 to 50 degrees Celsius idle and under about 85 degrees Celsius under load for the GPU. The hard ceiling, called the Tjmax, is 105 degrees Celsius on Intel Core Ultra and 95 degrees Celsius on AMD Ryzen 9000, and a chip that reaches its limit lowers clock speed to protect itself. That is thermal throttling, and it means cooling needs attention.
What Do You Need to Monitor CPU and GPU Temperatures?
You need a monitoring tool that reads the hardware sensors and a sense of the safe ranges before you load the system. Every CPU and GPU has temperature sensors on the die; the tool reads them and reports the number. The items below cover detailed reading, an in-game overlay, a quick check, and a way to apply load.
- HWiNFO reads every CPU and GPU sensor in detail, including package, per-core, hotspot, and memory junction temperatures, and logs them to a file.
- MSI Afterburner with RivaTuner Statistics Server draws an in-game on-screen overlay so you can watch temperatures during real play.
- Core Temp reads per-core CPU temperatures in a small window for a fast check, using very little memory.
- Windows Task Manager shows the GPU temperature on its Performance tab with nothing to install, though it does not show CPU temperature.
- A stress test or benchmark, such as Cinebench for the CPU or a GPU benchmark loop, applies sustained load so the peak temperatures appear.
Monitoring is the first step behind several other jobs. The guide to overclocking your GPU depends on keeping the core under about 85 degrees Celsius while you raise clocks, and a high reading under load points to the guide to diagnosing PC overheating. If dust is the cause, the guide to cleaning your PC usually drops the numbers several degrees.
What Are CPU and GPU Temperatures?
CPU and GPU temperatures are the readings from sensors inside the processor and graphics card that measure how hot the silicon runs, expressed in degrees Celsius. Both chips produce heat as they switch transistors, and the built-in sensors report that heat to monitoring software. Knowing which sensor you are reading matters, because several of them run hotter than the core on purpose:
- CPU package temperature reports the hottest reading across the processor and serves as the overall CPU temperature.
- CPU core temperatures report each individual core, which rise and fall as load moves between cores.
- GPU core temperature reports the main graphics processor reading and serves as the overall GPU temperature.
- GPU hotspot and memory junction temperatures report the hottest point on the die and the video memory, which read 10 to 20 degrees Celsius above the core and have their own higher limits.
The CPU package and GPU core are the numbers to judge against the safe ranges. The hotspot and memory junction are expected to be higher, so a memory junction near 100 degrees Celsius on a high-bandwidth memory card is within specification rather than a fault.
Which Tool Should You Use to Monitor Temperatures?
The tool decides which sensors you can see and whether you get an in-game overlay. Each reads the same hardware sensors but presents them differently, so the choice comes down to how much detail you want and whether you need a reading on screen during a game.
A complete setup pairs HWiNFO for detail and logging with MSI Afterburner for the overlay. Core Temp and Task Manager are best kept for a quick glance, and running every tool at once can cause sensor conflicts, so close the ones you are not using.
| Tool | Best For | CPU Temp | GPU Temp | In-Game Overlay |
|---|---|---|---|---|
| HWiNFO | Full sensor detail and logging | Yes | Yes | Through RTSS |
| MSI Afterburner + RTSS | In-game overlay | Yes | Yes | Yes |
| Core Temp | Quick per-core CPU check | Yes | No | No |
| Task Manager | Glance at GPU, no install | No | Yes | No |
How Do You Set Up an In-Game Temperature Overlay?
An on-screen overlay shows CPU and GPU temperatures over the game so you can watch them under real load without leaving play. The MSI Afterburner and RivaTuner Statistics Server pair draws the readings in a screen corner. Set it up in order:
- Install Afterburner and RTSS. Download MSI Afterburner from msi.com and keep RivaTuner Statistics Server checked in the installer. A reboot helps the monitoring driver hook into games.
- Open the Monitoring tab. Click the gear icon in Afterburner, then open the Monitoring tab where every sensor is listed.
- Add the sensors to the display. Select CPU temperature and GPU temperature and tick “Show in On-Screen Display” for each, then add GPU usage and frame rate if you want them.
- Set the overlay in RTSS. Open RivaTuner Statistics Server from the system tray, confirm “Show On-Screen Display” is on, and set the position, size, and colour.
- Launch a game to confirm. Start a game and check the overlay appears in the corner. If an anti-cheat blocks it, use the game’s built-in counter or the GPU vendor overlay instead.
With temperature, usage, and frame rate on screen together, a frame rate drop that lines up with a temperature spike points straight at throttling during that session.
What Are Safe CPU and GPU Temperatures?
A safe CPU sits at 30 to 45 degrees Celsius idle and 65 to 85 degrees Celsius under load, and a safe GPU sits at 30 to 50 degrees Celsius idle and under about 85 degrees Celsius under load. Each chip has a hard ceiling, the Tjmax, where it begins to throttle. The table sets the safe, warm, and throttling bands so any reading has context:
| State | CPU (package) | GPU (core) | What it means |
|---|---|---|---|
| Safe idle | 30–45°C | 30–50°C | Healthy at rest; a high idle points to airflow or paste. |
| Normal load | 65–85°C | 65–85°C | Expected under gaming or rendering; full clocks held. |
| Warm | 85–95°C | 85–90°C | Within limits but near the edge; watch cooling. |
| Throttling | 95–105°C (Tjmax) | ~88°C hotspot+ | Chip lowers clock speed to protect itself. |
The Tjmax differs by maker. Intel Core Ultra (Arrow Lake) is rated to 105 degrees Celsius and AMD Ryzen 9000 to 95 degrees Celsius, and both are designed to run at that limit safely rather than break. Holding the CPU under 85 degrees Celsius and the GPU under about 85 degrees Celsius keeps clear of the throttle point and leaves room for the headroom the GPU overclocking guide uses.
How Do You Spot Thermal Throttling?
You spot thermal throttling by watching clock speed and temperature together and catching the clock drop at the moment temperature hits the limit. A single temperature reading is not enough; the proof is in how the clock responds. Confirm it in order:
- Log temperature and clock during a stress test. Open HWiNFO in sensors-only mode, start logging to a file, and run Cinebench for the CPU or a GPU benchmark loop for 15 to 30 minutes.
- Watch for the clock dropping below its rated boost. A clock that falls and fluctuates the instant temperature peaks is the signature of throttling.
- Check the throttling flag. HWiNFO reports a “Thermal Throttling” sensor that reads Yes when the CPU is limiting itself.
- Tie it to frame rate. If frame rate sags 10 to 20 minutes into a session as temperature climbs, the heat is capping performance.
Once throttling is confirmed, the fix is cooling, not settings. Better airflow, fresh thermal paste, or clearing dust usually restores the lost clocks, and the overheating diagnosis guide works through the causes in order.
Last Thoughts on Monitoring CPU and GPU Temps
Monitoring CPU and GPU temperatures comes down to picking the right tool and reading the right sensor. HWiNFO gives full detail and logging, MSI Afterburner with RivaTuner Statistics Server puts a live overlay in your game, and Core Temp or Task Manager cover a quick glance. Judge the CPU package and GPU core against the safe ranges, treat the hotspot and memory junction as higher by design, and remember the Tjmax is 105 degrees Celsius on Intel Core Ultra and 95 degrees Celsius on AMD Ryzen 9000.
A chip that reaches its limit and loses clock speed is thermal throttling, and the answer is better cooling rather than lower settings. The readings here feed into overclocking, overheating fixes, and frame rate work across the rest of the PC tutorials hub.
Key Takeaways:
- Use HWiNFO for full detail and logging, and MSI Afterburner with RTSS for an in-game overlay; Core Temp and Task Manager cover a quick check.
- Task Manager shows GPU temperature but not CPU temperature, so use Core Temp or HWiNFO for the processor.
- Read the CPU package and GPU core as the overall temperatures, and treat hotspot and memory junction as 10 to 20 degrees Celsius higher by design.
- Keep the CPU at 30 to 45 degrees Celsius idle and under 85 under load, and the GPU at 30 to 50 idle and under about 85 under load.
- The Tjmax is 105 degrees Celsius on Intel Core Ultra and 95 degrees Celsius on AMD Ryzen 9000; at the ceiling the chip throttles.
- A chip that hits its limit and loses clock speed is thermal throttling and needs better cooling, not lower settings.
Frequently Asked Questions (FAQs)
How do I monitor CPU and GPU temperatures?
Install HWiNFO for full sensor detail, or MSI Afterburner with the RivaTuner Statistics Server overlay to see readings in-game. For a quick look, Core Temp shows per-core CPU temperatures and Windows Task Manager shows the GPU temperature on its Performance tab. Read the CPU package and GPU core values, then log them during a stress test.
What is a safe CPU temperature for gaming?
A CPU is safe at 30 to 45 degrees Celsius idle and 65 to 85 degrees Celsius under gaming load. The hard ceiling is the Tjmax, which is 105 degrees Celsius on Intel Core Ultra and 95 degrees Celsius on AMD Ryzen 9000. At the ceiling the chip throttles to protect itself, so staying under 85 degrees Celsius leaves headroom.
What is a safe GPU temperature?
A GPU is safe under about 85 degrees Celsius under load and 30 to 50 degrees Celsius at idle. Modern cards are rated to roughly 90 degrees Celsius, but performance is best below 85. The hotspot and memory junction sensors read 10 to 20 degrees Celsius higher than the core by design and have their own limits.
How can I tell if my PC is thermal throttling?
Watch the clock speed and the temperature together in HWiNFO or MSI Afterburner. Thermal throttling shows as the CPU or GPU clock dropping below its rated boost the moment temperature hits the limit, usually with frame rate falling after 10 to 20 minutes of load. HWiNFO also reports a “Thermal Throttling” flag that reads Yes when it happens.
Does Task Manager show CPU and GPU temperature?
Windows Task Manager shows GPU temperature on the Performance tab next to the GPU graph, but it does not show CPU temperature. For CPU temperature use Core Temp, HWiNFO, or HWMonitor. Task Manager is the fastest way to glance at the GPU without installing anything.
Why is my GPU memory junction so hot?
The memory junction sensor on high-bandwidth memory cards reads hotter than the core and stays within specification up to about 95 to 105 degrees Celsius. A reading near 100 degrees Celsius on the memory junction is normal, so compare it to its own limit rather than the core’s 85 degree threshold.
