How to Check PC Temperatures (Safe Ranges and Tools)

Learn how to check your CPU, GPU, and SSD temperatures in Windows and BIOS/UEFI, what “safe” ranges typically look like, and which monitoring tools are worth using (plus how to interpret confusing readings like CPU “Tj”/

TL;DR

• Read your CPU package/core temps, GPU temp/hotspot, and SSD temps first with a trusted monitor (HWiNFO, HWMonitor).
• Want to see if something looks strange? Monitor temperatures at idle (usually 10 minutes after log-in for Windows) and at load (10-20 minutes of your real heaviest workload).
• When the article says “Safe” it typically means stay below the published maximum for that component (usually Tjmax/Tjunction max, but not always) and to ignore the chase for one number to rule them all.
• Windows Task Manager will show a GPU temperature on most systems, though Windows will usually not show a CPU temperature by default.
• If you’re struggling with cooling and are close to the threshold if not throttling, consider fixing things like airflow, dust, fan curves, or cooler mounting before overclocking or otherwise trying tweaks.

Why check PC temps

PC components all have temperature limits, usually a range around which hovering for long periods leads to thermal throttling (lowering clock speeds), or in the worst cases, thermal shutdown (powering off). Checking temps frequently helps catch those that glow too hot (to protect themselves, of course) and can be a first alert to trouble like dust inside, a bad fan or supply of air, dried thermal paste, or other thermal issues. You can also verify new cooling components, hardware changes, or other configuration changes are delivering the improvement you might expect.

Understand the numbers you see

Temperatures for “Core” vs Package vs the “real limit” (Tjmax/Tjunction max)
CPU temps are often a point of confusion, with CPU monitor programs showing multiple temperatures. “CPU Package” (or something similar) is often the most useful single number for overall CPU heat under load, whereas “Core” temperatures can let you know if one part of the chip runs significantly hotter than others.

For you, what matters most is your CPU’s maximum allowed operating/junction temperature (sometimes giving the name Tjunction max or Tjmax, and similar). Intel specifically states that Tjunction max denotes the temperature threshold at which the CPU will begin to utilize its built-in thermal controls (such as throttling) in order to reduce temperature, and this limit was different for each Intel CPU model. It is commonly in the range of 100-110°C for many Intel CPUs.

GPU temperatures: “GPU temperature” vs “hotspot/junction”
GPUs will often report more than one temperature, and primarily you will likely see a basic “GPU Temperature” and then a few other metrics like “Hotspot” (sometimes known as “Junction”). Hotspot/junction is the hottest of either all or a sample of the sensor areas on the die, meaning it can look quite alarming even when the GPU is behaving normally. A reason for why labels for these sensors matter a lot is that AMD states (for some but not all Radeon designs/architectures) that a 110°C junction/hotspot reading can be within spec, and the average/edge temperature is lower. So don’t compare your “hotspot” reading with someone else’s GPU temp screenshot—they are not using the same metric.

SSD temps and motherboard temps: don’t overlook them
NVMe SSDs can throttle when hot, which can lead to lower sustained write speeds. Your motherboard’s VRMs (voltage regulator modules, aka power delivery) can also get hot, especially in compact cases with bad airflow. Good monitoring tools can expose these so you can improve case airflow before you start to experience instability.

Typical “safe” ranges (and what should ring alarm bells)

There isn’t one safe temperature, because the actual limit is model dependent (Tjmax/Tjunction max). What follows are ballpark numbers that should help with sanity checking – particularly if you’re also on the lookout for performance drops, noisy fans, stuttering, crashes, or sudden reductions in clock-speed (signs of throttling).

Ways to check PC temperatures (Windows + BIOS/UEFI)

Method 1: BIOS/UEFI hardware monitor (best for a quick CPU check)

• Most motherboards have a handy hardware monitor screen in BIOS/UEFI that shows you how hot your CPU is and how fast your fans are spinning. This is useful if you’ve just put together a new PC, installed a CPU cooler, or gig a fan or pump that’s not running.
• Just don’t expect these BIOS temps to be a perfect “idle baseline”—your CPU isn’t in quite the low-power state it gets in Windows.
• 1. Reboot your PC and enter BIOS/UEFI (make note of how to get in specific to your motherboard. Del and F2 are the most common).
• 2. Find a menu that says something like Hardware Monitor / Fan Control / PC Health.
• 3. Read off CPU temp and make sure CPU fan and (if it’s in one of these menus) your AIO pump’s RPM are reporting sensibly.
• 4. If CPU temp climbs rapidly at idle in BIOS, shut down and re-check your cooler mounting and thermal paste.

Method 2: Windows Task Manager (built-in, but GPU-only on many PCs)

• On many Windows 10/11 PCs and laptops, you can get a quick GPU temperature readout in Task Manager. On lots, there isn’t a built-in CPU temperature readout in Task Manager—but on those CAN include temperature monitoring features, it’s a simple tool.
• 1. Press Ctrl + Shift + Esc to open Task Manager directly.
• 2. Click on the Performance tab.
• 3. Click GPU or GPU 0 / GPU 1 if you have more than one GPU, to see your current GPU temp (shown in °C). If you don’t see GPU temperature, update your graphics driver and make sure your system meets Task Manager GPU temperature criteria (some systems require a newer WDDM model driver).

Method 3 (recommended): Dedicated monitoring tools (CPU + GPU + SSD)

• For most, a dedicated hardware monitor is the simplest way to obtain accurate, detailed temps (and log them over time). These tools can also report “max” values—important because spikes can happen quickly and be missed in real time.
• HWiNFO (excellent sensor coverage; easy to read with a Sensors-only mode).
• HWMonitor by CPUID (quick and simple overview of core temps, GPU temp, SSD temps, voltages, etc.).
• Open Hardware Monitor (open-source; can show temps, fan speeds, clocks, and load).

Method 4: Vendor tools (nice for “official” telemetry and tuning).

• AMD Ryzen Master: Shows temperature, clocks, voltages, and used for monitoring (and tuning) on many supported Ryzen CPUs.
• Intel Extreme Tuning Utility (Intel XTU) is an overclock/monitor/stress tool for supported unlocked Intel processors (Intel provides a different Intel XTU for each CPU generation).
• MSI Afterburner, popular for GPU monitoring and displaying an overlay while gaming (even if you’re not an overclock!)

Step-by-step: temps in detail (idle, then load)

1. Pick one monitoring utility—HWiNFO or HWMonitor—and run it with admin rights if it prompts.
2. Note the temperature in your room; referred to as ambient. This strongly affects results (especially if you’re in a small room, or a warm climate), so do note!
3. Idle: after Windows starts, do nothing for around 10 minutes. Write down CPU package temp, GPU temp, SSD temp, and fledgling speeds (fan speeds).
4. Real workload: play a game you actually play, do a render/export you do do, or run your normal “heaviest” app for ten or twenty minutes. Note average and max temps.
5. Stress test (only if you know what you’re doing, your stress test may run hotter than normal use and be no way indicative of normal temps): see if your monitoring tool has a logging feature, use that and look through the log after a normal session (max temps can be easier to trust than “glancing” at the screen), and also check for throttling indicators (drops in frequency, its power-limiting functionality kicking in, sudden drops in FPS).
6. Compare your max readings with the limit for the component—you want Tjmax/Tjunction max. Watch out for throttling indicators (such as the CPU dropping in frequency, its power-limiting feature kicking in, or a sudden drop in FPS).

How to find your component’s real maximum temperature (so “safe” is not guesswork)

• Intel CPUs: look up Tjunction max (Tjmax) (Intel) recommend checking your CPU ‘exact make/model’ thermal specs on its Product Specifications (ARK) pages (Intel) — they explain Tjunction max here and stress that the value does vary per product, but many of them are in the ~100-110°C range).
• AMD CPUs: look for “Max Temps” / “Tjmax” on the CPU’s product page. AMD will often make the operating temp (Tjmax) available on the CPU’s product page, in the way that the AMD Ryzen 7 5800X product listing shows an operating temp (Tjmax) of 90°C. Use your exact model’s listing/spec page, not someone else’s limit unless you have the same CPU.

If your temps are high: practical fixes (in order from least effective to most)

No-tools checklist (five minutes or less)

• Make sure your PC has adequate space around it; if it’s pressed against a wall or in a closed cabinet it won’t be able to effectively vent heat away from itself.
• Clean dust from dust filters and vents. Dust is a primary reason for gradually rising temps over time.
• Ensure all case fans are spinning freely and oriented properly (front/side intake, rear/top exhaust is a common layout).
• Laptop users: use a hard surface. Soft surfaces may block intake airflow vents.

Software / settings checks

• Check out any background CPU usage (launcher updates, browser tabs, etc) that might prevent your CPU from truly idling.
• Update your BIOS/UEFI and ensure you’ve installed the motherboard chipset drivers if you’re troubleshooting a new build or odd fan control behavior.
• Look through the fan curves in BIOS or your motherboard utility, if it has one. An overly “quiet” fan curve may allow heat to spike unnecessarily.
• If you’ve overclocked the CPU or changed power limit settings, reverting back to stock is a good way to establish a baseline first.

Hardware fix (more impact, more effort)

• Re-mount the CPU cooler. Uneven mounting pressure is a common reason for high CPU temps.
• Replace the thermal paste: if it’s old or there’s too much/too little of it on the CPU cooler it can raise your temps. Replace it if it’s been years or you suspect you have a bad CPU cooler mount.
• AIO liquid coolers: make sure the pump is running, and also that the fans on the radiator are ramping up under load.
• Case airflow: adding fans to a case or replacing old case fans with a higher performance option drives your CPU and GPU temps lower simply because of lowering the ambient temperature of the case.

Common mistakes while checking for temperatures

• Comparing readings on different sensors (cpu core and package temp; gpu cool and hotspot/junction) and assuming one number is “wrong” if they don’t match.
• Only judging temps by idle. Many cooling problems don’t show up until there’s been a sustained load for 10–20 mins.
• Using the BIOS/UEFI temperature reading as your base for Windows idle (they’re not in the same “state” of operation).
• Installing monitoring tools from unofficial mirrors (more dangerous of getting bundled adware or worse).
• Chasing lower temperatures for the sake of it, at the cost of noise or performance, and without having an end state in mind (the real goal here is stability and staying within spec).

FAQ