battery — The CPU core temperature is over 110 °C. Is it too high?
2 years, 1 month ago
Are these temperature values below the permissible level?
The max temperature for many CPUs is listed in the 105-110°C range.
But for long-term use, you’re much better off keeping things below
80°C in general and only pushing up to 85°C at the most
To make sure your software is giving you an accurate reading you could compare the results with those of another temperature monitoring application. There are several free lightweight tools to monitor your CPU temperature, one of which is Core Temp.
It would be advisable to check your heat sink to see if it has been dislodged, or that the thermal paste is still connecting the heat sink to the CPU, as running your CPU at that temp is sure to cause damage in the long run.
The GPU temperatures that are shown in your question are not in the danger zone.
You’re well into the range that can cause permanent damage to the CPU. Most consumer electronics (including CPUs) are not designed to operate above 85 °C for any extended period of time, and most will actually shut down when they get over about 100-105 °C. Provided you have a working (and properly sized) cooling system and are not somewhere with unusually high temperatures to begin with (40 °C or higher), you should not be seeing temperatures that high no matter how hard you push the CPU.
However, I’m inclined to believe something is wrong with your system due to that insane discrepancy between reported package and core temperatures. In particular, the possibilities that come to mind are:
- The sensor isn’t being read correctly and the core temperature is actually much lower. This is the best possible case, and it’s easy to check (try a handful of other tools for reading these sensors. Everything else reporting similar temperatures does not rule this out though, because the drivers being used to make the reading may be bad (you can check that case by booting into a live Linux environment and seeing what it says the temperatures are. I know 100% for certain that Linux reports the temperature correctly on that model of CPU because the sensor interface the CPU provides has been around since the AMD K10 days and is very well supported by Linux).
- The sensor isn’t being read correctly, and the package temperature is actually much higher. This is extremely unlikely, because for it to be the case you have to have somehow managed to run a CPU with a 20 W TDP so hard that it got that hot. The only possibilities I can think of that would allow for that are running with no cooling system at all or running in an environment that was already unlivably hot for humans.
- Something is wrong with one of the temperature sensors. It is not very likely, but it is still possible. If you’ve eliminated the two above possibilities, then this one can be checked by using a (good) infrared thermometer or (real) thermal camera to get an estimate of the temperature of the junction between the heat-sink and the IHS. If that reads back at close to 115 °C, then the package temperature sensor is bad (and something else is probably wrong with your CPU). If it reads back at close to the 66 °C being reported by the package temperature sensor, then either something is wrong with the core temperature sensor or the next (worst case) possibility is the case.
- Something is physically wrong with the thermal junction inside the package between the IHS and the CPU die. This is the absolute worst case scenario, as it means your chip is essentially useless (because you quite simply cannot cool it well enough for it to be practically usable. This is also astronomically unlikely (actually, it’s beyond astronomically unlikely, it’s even beyond the unlikelihood of a SHA-256 hash collision with two randomly chosen files), but it’s still technically possible. There’s unfortunately no practical way to check this one if you’ve eliminated all the other possibilities, because delidding the CPU to manually check will make it irrelevant (and also require use of a completely different cooling system).
CPU core temperatures of more than 110 degrees is too high and make the processor stop working. If really temp is over 110 degrees system will likely crash and there could be nasty situation. HWMonitor is showing incorrect values.
Go to your BIOS and check temperatures there, and if BIOS also says nothing this could be indicative that your Motherboard dosen’t have any temperature sensors, thus explaining why HWMonitor is showing strange values.
If the temperature is high, then you can check the airflow, clean the dust in casing and components and check if the fan is moving properly. You can use Core Temp as the other answer suggested.
Your CPU temperature is reaching its very maximum. I think, any PC component having temperature over 100C is too much for extended period of time. It is recommended to keep it under 80C in the long run, 60-something is even better.
I always recommend to remove the original cooling fan of the CPU (which usually give you when you buy it) and buy a better cooler with bigger heatsink. I bought CoolerMaster HYPER TX3i, which is a cheap (under $30), but an effective one.
For my i3-9100F temperature with the original heatsink cooler: Idle — 65C, Load — 95 — 100C; after applying CoolerMaster: Idle — 30 — 42C, Load — 50 — 60C.
When a CPU reaches 100+ degrees, it should be restarted. If it still doesn’t’ resolve the issue, you might want to reset your PC or delete some apps. Another reason might be the heatsink or the cooler’s problem. If you don’t want to get new hardware, I recommend you turn the CPU voltage a bit lower. If I am correct, your CPU is not unlocked, so you can’t really adjust your clock speed, but beware when turning your voltage down, turn it down bit by bit. If you notice instabilities, revert the changes.
The AMD A10-7300 has a maximum operating temperature of 102°C so your system should be shutting itself down to prevent damage.
The iGPU (Radeon R6) temp of 66°C appears to be much more believable. Since this matches the «Package Temp» then I would surmise that your actual temp is 66°C since the CPU cores and iGPU are in the same housing.
Given the fact that your cores are all running well above the stock 1900 MHz and one of them running near the max 3200 MHz turbo, it tells me that the temperatures are not affecting performance. If you’re temperatures were too high then your CPU would throttle down to prevent damage.
You can download other temp monitoring software and see if you get similar results. Also, if your BIOS reports temps then you should check temps in your BIOS since that should be the most accurate reading.
The temperature you told and I am seeing from the software is actually not that high! Because it can happen if you play high-end games or video editing software and benchmark software.
Although it’s generally not an ideal temperature for a computer, but if the temperature stays while you do simple working and light or mid high works in your computer then it can be your cooling system. Which is failing!
You can use the software CPU-Z for testing your computer’s temperature and then reask the question. Because sometimes software can be faulty, but CPU-Z is one of the best!
The ideal temperature for a normal working computer is between 62-66 degree Celsius. Anything above can be bad for your motherboard and its components if not for the processor itself.
You can download CPU-Z from here: CPU-Z
Sign up or log in
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Post as a guest
Required, but never shown
One CPU Core Hotter Than Others (Causes, Fixes)
If you overclock your CPU or frequently run CPU-intensive applications, you may have noticed abnormal heating inside the CPU.
Using temperature-monitoring apps can help you keep these temps in check and make sure nothing goes above the normal threshold.
However, as these apps show you the temperature of each CPU core, you may notice that not all cores have the same temperatures.
Is it normal, or should you be worried?
This article addresses everything you need to know about CPU core temperatures.
1. Normal CPU Temps
An increase in CPU temps while the computer is in use or even idling is normal as heat is a normal by-product of computing processes.
However, there’s a certain limit to how hot your CPU cores can get because overheating can create serious problems for your PC or laptop.
When it comes to determining the normal CPU temperature range, there’s no one-size-fits-all answer since this range depends on many different factors, such as your computer’s make, components, and applications that you run at a given time.
For example, when your computer is running idle, it will have different temperatures than when you run CPU-intensive applications like games or video-editing tools.
While you should check your specific CPU brand’s normal temperature range, there’s an overall safe range to keep your CPU temps within it.
The best idling temperature for most CPUs is 40–45°C and 65–75°C while running essential apps.
However, the CPU can reach up to 90°C while playing games and running heavy programs.
In addition, laptops can run at higher temperatures because their components are more densely packed and have lower airflow.
When Should You Worry?
Although computers frequently run hot, especially under full load, there’s little to worry about.
High CPU temperatures don’t damage your CPU or motherboard because your computer has a failsafe mechanism that turns it off when it gets too hot.
That’s a helpful sign that tells you there’s something wrong with your PC components and you should take care of it.
If the computer turns off or restarts randomly or it keeps crashing under normal applications, you need to fix the issue.
That’s particularly urgent if you overclock and experience temperatures higher than 90°C since high temperatures can kill your CPU if the voltage also goes up.
2. Normal Core Temp Differences
It’s important to monitor the temperature of your CPU and each core to make sure everything is running smoothly and trouble-free.
However, as you check your CPU temps using third-party tools such as Core Temp, you may notice that each core has different temperature readings.
That’s an issue frequently reported by users who want to know if this difference in temperatures is normal or a cause for concern.
The main reason one core is running hotter than others is the way most applications work.
Most of the programs we run on computers are single-thread, meaning they only need one core to perform their tasks.
As a result, they occupy one core, which is typically the first core, and others run idle or perform heavier tasks.
One core is always working more than others.
This higher workload leads to higher temperatures, making the core different from others.
You may even notice that one of the other cores (and not necessarily the first one) is hotter because systems are different, and they may assign programs to different threads.
In addition, multi-core CPUs are designed so that one core is surrounded by others, making that core hotter than others due to lower airflow.
However, not any core temp difference is normal, and you should be worried when cores are hotter than others for more than 10 degrees.
In such cases, you should look for the possible causes of these high temperatures.
One of the primary culprits is thermal paste.
It could be old, not properly applied, or even not good quality.
The thermal paste has to be spread evenly over the heatsink, and even microscopic air bubbles under the thermal paste can cause temps to rise.
Make sure all parts are covered and the heatsink is installed correctly.
3. Reapply Thermal Paste
While there are different ways of applying thermal paste, one of the best ways, which is also the simplest, is the dot method.
You place a pea-sized dot of paste on the CPU and put the cooler on it.
Some people recommend using the line method, which means applying a line of paste on the CPU, but this can lead to less favorable results as it may not spread evenly.
Others prefer to go by the manual method, using a flat object to spread the paste because it gives you a smooth cover and helps you control the amount of paste.
However, it’s not recommended because it can create air bubbles.
Applying enough thermal paste can be tricky because today’s CPUs are bigger and may need more of that.
If you apply too little, you won’t get good results in cooling down the CPU, and applying too much will lead to messy spills.
In addition, if the thermal paste is conductive, you’ll get electric shorts and damage your motherboard.
Before applying thermal paste, clean the CPU and the heatsink with a microfiber cloth and ensure there’s no lint or old thermal paste left on the surfaces using isopropyl alcohol.
Placing the cooler is the trickiest part because you have to get it right on the first attempt.
If you don’t put it in the right place and direction, you’ll need to remove it, increasing the chances of getting air bubbles and an uneven spread.
Another thing to remember is that you should apply enough force while pressing the cooler down on the CPU.
The pressure should be small enough to avoid damaging the CPU and motherboard and big enough to get an even and smooth spread and prevent the cooler from sliding.
After applying the paste, inspect all components and ensure there’s no spill.
If you see thermal paste around the CPU, you’ve applied too much thermal paste.
In that case, you should clean off the paste and redo the entire process.
4. Delid The CPU
If reapplying thermal paste doesn’t work, you may need to take a more drastic measure, which can be scary as it’s delicate and needs a high skill level.
Delidding the CPU is a process used only by elite overlockers, but today’s delidding kits have made it much easier.
The CPU has a “lid” known as the Integrated Heat Spreader (IHS).
The heat sink on the CPU sends the heat from the CPU to the cooler.
Delidding means removing this lid and replacing thermal paste or liquid metal with more effective materials to better cool down the CPU cores than the stock thermal paste did.
Since the process is sensitive, make sure you know what you should do before and go over your CPU and thermal paste’s user manual to avoid possible mistakes.
Here’s a helpful video that shows you how to use the delidding kit components to delid your CPU.
5. Close Background Programs
While the CPU cores running hot may seem like a hardware-related issue, certain software programs can also lead to heating.
As mentioned, some CPU cores work harder than others, making them hotter.
You can lower their temperature by reducing their workload.
CPU-intensive programs running in the background can also cause CPU heating issues, including one core running hotter than others.
You could overlook the possibility of this simple cause because you may think the problem is more serious than that.
However, before trying other hardware-related methods, it’s recommended to close all background applications and monitor the temps to see if it changes anything.
Go to the Task Manager by pressing Alt + Ctrl + Del and check the Processes tab.
Go over the programs and see if you can close any of them, especially those taking up many CPU resources.
Now, run your temperature-monitoring program and see if it helps.
Another thing that can lead to CPU core heating is many temporary files containing useless data and junk files.
These files place a huge burden on the CPU cores and eat up lots of storage space.
Use an effective cleanup program to ensure your computer is junk-free.
6. Check For Faulty Sensors
Getting different temperature readings for different cores is generally a normal thing because of the way the CPU thermal sensors work.
Most of these sensors aren’t designed to give accurate readings, and their purpose is just to avoid thermal overload.
As a result, they don’t read temperatures below a certain value (around 50°C) and only give you a general indication of the CPU’s thermal state.
If the readings are different for up to 10°C, there’s nothing to worry about.
However, if you get large differences in readings and reapplying thermal paste and delidding doesn’t help eliminate the issue, you should ensure the CPU thermal sensors give you the correct readings.
That said, locating these sensors isn’t easy because every CPU manufacturer places them in a different location.
They may be on the CPU die or the interposer and you probably need to search and ask informed technicians to make sure your sensors are healthy.
7. Try General Cooling Methods
The temperature issues inside the CPU cores aren’t different from other thermal problems across the computer.
So, you can take all the measures for cooling down your PC and see if it helps.
Here’s what you can do:
A. Keep Your PC Clean
Dust buildup is one of the most important causes of overheating.
When dust accumulates on computer components, it can work as an insulator and prevent airflow inside the computer.
You need to regularly clean your PC, especially if it’s in a place with lots of dust.
Try cleaning your PC and see if the core temperature difference goes away.
Make sure to clean every nook and cranny, particularly the fans.
Be careful not to damage the sensitive components and use cotton swabs to clean tiny and hard-to-reach areas.
You could also use canned air to drive out the dust from the areas you can’t see.
By cleaning your PC, you’ll not only address the temperature issue, but you’ll also improve your PC’s performance and speed.
B. Improve Airflow
Although computers normally do a good job of keeping their components cool, you need to help your PC perform better.
One of the best things you can do is make sure there’s enough airflow going in and out of the PC.
Avoid putting your case inside cabinets or on soft surfaces like cushions or beds.
You want to make sure the air vents in different parts of the case aren’t blocked as it can severely affect the airflow.
However, you should avoid placing your computer in areas that get a lot of air, such as in front of the window or air-conditioner.
Plus, although you may think your laptop can be fine sitting on your lap, it’s not the optimal place for it because your legs can block airflow.
Place it on a desk, table, or a cooling pad with working fans.
If your computer is in a room with poor ventilation, you should make sure cool air can circulate easily around the computer by turning on the AC.
And never put your computer in a place where it gets sunlight for too long because it can heat the computer no matter how much airflow it has.
C. Check All The Fans
The fans in both desktop and laptop computers are essential in keeping the internal parts cool.
You should always make sure they’re operating normally and keep them dust-free.
Since fans have moving parts, they can wear down over time, reducing the fan’s efficiency in cooling down the innards.
Check the fans for any physical damage, loose wires, or broken blades, particularly for unusually loud noises.
If you can’t fix the issue, you should replace the fan.
Another thing to consider is the filters on the fans if there are any.
Ensure the filters are clean and not damaged, and replace them if they don’t work properly.
D. Stay Updated
Here’s another software-related issue that can cause one core to work harder and get hot.
When your programs are outdated, they don’t run as smoothly as they should and place a lot of burden on the CPU by eating up its resources.
Look for regular Windows updates and download them automatically or manually to ensure your system is updated and free of glitches.
Another factor that can cause programs to eat up CPU resources is malware and viruses.
Windows built-in defender normally does a good job of protecting your PC against viruses.
Still, you could install a reliable antimalware and antivirus tool to prevent your PC from getting infected and remove unwanted or harmful programs.
E. Change The CPU Cooler
If the CPU cooler has lost its efficiency in cooling down the system, you can replace it with more powerful ones.
Depending on your system specs and needs, you could go for air-cooling or liquid-cooling solutions.
Liquid-cooling solutions use a liquid, which is typically a glycol solution, to circulate the system and cool it down.
Some of these cooling solutions have light indicators that show you the thermal state of the CPU and the entire PC, but you need glass side panels on your case.
You can choose among different designs, including closed-loop and open-loop consisting of a heatsink, tubes, a radiator.
The best thing about these solutions is that they don’t require much maintenance as they’re sealed and very effective.
They dissipate heat much better than their air-cooling counterparts and are much quieter because they don’t have fans like those.
However, they’re more expensive and can be prone to leakage.
On the other hand, air-cooling units merely circulate the air inside the PC using a heatsink and fan that dissipate heat from the CPU to other parts and outside.
Most stock coolers are of this type, and you can replace them with more robust and effective options to improve cooling.
These units are simpler, so you can easily upgrade them without changing the tubing or the radiator.
In addition, you can use them on almost every system since liquid cooling solutions need specific designs that not every computer has.
However, these units can lead to higher temps inside the case because they dissipate heat from the CPU inside the case.
Find Out CPU Temperature from the Command-Line
In this tutorial, we’ll take a look at how we can check for the CPU temperature on the Linux terminal. First, we’ll see how we can figure out the temperature of our CPUs without the help of third-party tools. Afterward, we’ll cover a couple of small useful utilities for the same purpose.
2. Finding CPU Temperature Without Third-Party Tools
On Linux, we can read almost every accessible detail related to hardware resources. These details include the count of the CPU cycles, CPU temperature, I/O usage, network usage, and more. This is all possible because Linux gives us more control over the hardware and software.
The /sys directory is a virtual file system that contains a plethora of information regarding the Linux kernel and the hardware. The files inside this directory don’t actually reside on the disk. Instead, they’re only created and updated on-the-fly as we read them.
The /sys/class directory is the hierarchy of the hardware. This directory mostly contains information about the devices that are registered with the kernel.
One of the directories is called thermal, which contains temperature information of the hardware resources:
$ ls -lL /sys/class/thermal total 0 . .. drwxr-xr-x 3 root root 0 Apr 7 00:05 cooling_device7 drwxr-xr-x 3 root root 0 Apr 7 00:05 cooling_device8 drwxr-xr-x 4 root root 0 Apr 7 00:05 thermal_zone0 drwxr-xr-x 3 root root 0 Apr 7 00:05 thermal_zone1 drwxr-xr-x 3 root root 0 Apr 7 00:05 thermal_zone2
In this directory, we’re concerned with the thermal_zone directories. The thermal_zone directories correspond to the thermometers placed on our motherboard.
Let’s cd into the thermal_zone0 directory and check what it contains:
$ cd thermal_zone0 && ls -lL total 0 ... drwxr-xr-x 2 root root 0 Apr 7 00:05 subsystem -rw-r--r-- 1 root root 4096 Apr 7 00:55 sustainable_power -r--r--r-- 1 root root 4096 Apr 7 00:29 temp -r--r--r-- 1 root root 4096 Apr 7 00:29 type -r--r--r-- 1 root root 4096 Apr 7 00:29 trip_point_0_temp ...
As we can see, it includes lots of files and directories. However, we’re only interested in the temp and the type files.
The temp file contains the actual temperature of the zone. It should contain just a single integer value:
$ cat temp 27800
We can divide this value by 100 to get the actual temperature in Celsius. In this case, it would be 27.8 °C.
The type file contains a value that signifies the zone to which the temperature corresponds:
$ cat type acpitz
The acpitz thermometer is located beside the CPU socket. However, we are interested in the CPU temperature. Similarly, for the CPUs, we can check the other thermal_zone directories that might contain this thermal information.
2.4. Putting It All Together
It can be tedious to check for CPU temperature this way because these directories might be different on different machines. Interestingly, the zone information is defined in the driver for the hardware resources.
For that reason, we might want to use a command that prints out this information in a readable way:
$ paste <(cat /sys/class/thermal/thermal_zone*/type) <(cat /sys/class/thermal/thermal_zone*/temp) | column -s $'\t' -t | sed 's/\(.\)..$/.\1°C/' acpitz 27.8°C acpitz 29.8°C x86_pkg_temp 38.0°C
Let’s break it down:
- We read the type and file files from each thermal_zone directory and feed the result to paste
- The paste command will align the lines from the corresponding files, separated by tabs
- We pipe the output of the paste command to column, which further aligns the output into columns
- The contents of column are then piped to sed, which replaces the values with readable temperature values
Additionally, we can create a simple shell script out of this rather long command and execute it either directly or from another script.
lm_sensors is a handy utility for monitoring temperatures, voltage, fan speed, and other hardware sensor information.
On major Linux distributions, lm_sensors should already be installed. However, if it’s not, we can use a package manager to install it from our distro’s official package repository:
# Ubuntu-like $ apt install lm-sensors
# Fedora, RHEL, openSUSE $ yum install lm_sensors
# Arch-like $ pacman -S lm_sensors
Once installed, let’s verify it:
$ sensors -v sensors version 3.6.0+git with libsensors version 3.6.0+git
We can use lm_sensors by simply typing in the sensors command:
$ sensors acpitz-acpi-0 Adapter: ACPI interface temp1: +27.8 C temp2: +29.8 C coretemp-isa-0000 Adapter: ISA adapter Package id 0: +40. 0 C Core 0: +39.0 C Core 1: +40.0 C
As we can see, the CPU temperature for each core is given in the Core 0 and Core 1 fields, respectively.
Moreover, if it doesn’t display the CPU temperature, we can run the sensors-detect command beforehand. The sensors-detect command will detect all the available sensors attached to the machine.
acpi is another lightweight alternative that we can use to display the temperature and battery information.
The acpi utility doesn’t ship with most distributions, so we’ll have to install it from our official package repository using the package name acpi:
# Ubuntu-like $ apt install acpi
# Fedora, RHEL, openSUSE $ yum install acpi
# Arch-like $ pacman -S acpi
After the installation, let’s verify it:
$ acpi -v acpi 1. 7
We can print out the temperature information with acpi by simply running it with the -t or –thermal option:
$ acpi -t Thermal 0: ok, 29.8 degrees C Thermal 1: ok, 27.8 degrees C
We can print a detailed report with the -i or –details option as well.
In this article, we covered how we can check the thermal status of our CPUs. First, we experimented with the raw thermal details provided by the Kernel in the /sys/class directory. Afterward, we used a couple of alternative tools that automate this process for us.
If you have a few years of experience in the Linux ecosystem, and you’re interested in sharing that experience with the community, have a look at our Contribution Guidelines.
lm_sensors — ArchWiki
lm_sensors (Linux monitoring sensors) is a free and open-source application that provides tools and drivers for monitoring temperatures, voltage, and fans. This document explains how to install, configure, and use lm_sensors.
- 1 Installation
- 2 Setup
- 3 Running sensors
- 3.1 Adding DIMM Temperature sensors
- 3.2 Reading SPD values from memory modules (optional)
- 4 Using sensor data
- 4.1 Graphical front-ends
- 4.2 sensord
- 5 Tips and tricks
- 5.1 Adjusting values
- 5.1.1 Example 1. Adjusting temperature offsets
- 5.1.2 Example 2. Renaming labels
- 5.1.3 Example 3. Renumbering cores for multi-CPU systems
- 5.2 Automatic lm_sensors deployment
- 5.3 S.M.A.R.T. drive temperature
- 5.1 Adjusting values
- 6 Troubleshooting
- 6.1 K10Temp module
- 6.2 Asus B450M-A/A320M-K/A320M-K-BR motherboards
- 6.3 Asus B450/X399/X470 motherboards with AM4 Socket
- 6.4 Asus H97/Z97/Z170/Z370i/X570/B550 motherboards
- 6.5 Asrock Deskmini h570
- 6.6 Gigabyte B250/Z370/B450M/B560M/Z690 motherboards
- 6. 7 Gigabyte GA-J1900N-D3V
- 6.8 Laptop screen issues after running sensors-detect
- 6.9 i2c bus errors on AMD Navi 2 GPUs
Install the lm_sensors package.
Note: More documentation is at the GitHub repository. In the future these may be installed, see FS#48354.
Use sensors-detect as root to detect and generate a list of kernel modules:
Warning: Do not use anything other than the default options (by just hitting
Enter), unless you know exactly what you are doing. See #Laptop screen issues after running sensors-detect.
It will ask to probe for various hardware. The «safe» answers are the defaults, so just hitting
Enter to all the questions will generally not cause any problems. This will create the
/etc/conf.d/lm_sensors configuration file which is used by
lm_sensors. service to automatically load kernel modules on boot.
When the detection is finished, a summary of the probes is presented.
This program will help you determine which kernel modules you need to load to use lm_sensors most effectively. It is generally safe and recommended to accept the default answers to all questions, unless you know what you're doing. Some south bridges, CPUs or memory controllers contain embedded sensors. Do you want to scan for them? This is totally safe. (YES/no): Module cpuid loaded successfully. Silicon Integrated Systems SIS5595... No VIA VT82C686 Integrated Sensors... No VIA VT8231 Integrated Sensors... No AMD K8 thermal sensors... No AMD Family 10h thermal sensors... No ... Now follows a summary of the probes I have just done. Just press ENTER to continue: Driver `coretemp': * Chip `Intel digital thermal sensor' (confidence: 9) Driver `lm90': * Bus `SMBus nForce2 adapter at 4d00' Busdriver `i2c_nforce2', I2C address 0x4c Chip `Winbond W83L771AWG/ASG' (confidence: 6) Do you want to overwrite /etc/conf. d/lm_sensors? (YES/no): ln -s '/usr/lib/systemd/system/lm_sensors.service' '/etc/systemd/system/multi-user.target.wants/lm_sensors.service' Unloading i2c-dev... OK Unloading cpuid... OK
Note: A systemd service is automatically enabled if users answer YES when asked about generating
/etc/conf.d/lm_sensors. Answering YES also automatically starts the service.
coretemp-isa-0000 Adapter: ISA adapter Core 0: +35.0°C (crit = +105.0°C) Core 1: +32.0°C (crit = +105.0°C) w83l771-i2c-0-4c Adapter: SMBus nForce2 adapter at 4d00 temp1: +28.0°C (low = -40.0°C, high = +70.0°C) (crit = +85.0°C, hyst = +75.0°C) temp2: +37.4°C (low = -40.0°C, high = +70.0°C) (crit = +110.0°C, hyst = +100.0°C)
Adding DIMM Temperature sensors
This article or section needs language, wiki syntax or style improvements. See Help:Style for reference.
Reason: Some style issues. In particular, section should avoid first person (i.e. «In my …») and the language in some sentences can be improved for readability and compliance with Help:Style#Spelling and Help:Style#Language. (Discuss in Talk:Lm sensors)
To find the temperature sensors of DIMMs, install the i2c-tools package. Once installed, load the
i2c-dev kernel module.
# modprobe i2c_dev
To show all the columns, use i2cdetect as root:
# i2cdetect -l
i2c-2 smbus SMBus PIIX4 adapter port 2 at 0b00 SMBus adapter i2c-2 smbus SMBus PIIX4 adapter port 1 at 0b20 SMBus adapter i2c-0 smbus SMBus PIIX4 adapter port 0 at 0b00 SMBus adapter
Otherwise, its output will appear as follows:
i2c-2 unknown SMBus PIIX4 adapter port 2 at 0b00 N/A i2c-2 unknown SMBus PIIX4 adapter port 1 at 0b20 N/A i2c-0 unknown SMBus PIIX4 adapter port 0 at 0b00 N/A
In my system, RAM sticks connected to the bus is SMBus 0. The i2cdetect command will show the devices that connected to the bus. The
-y 0 argument means use i2c-0 smbus. You can check other buses if needed.
# i2cdetect -y 0
0 1 2 3 4 5 6 7 8 9 a b c d e f 00: -- -- -- -- 0c -- -- -- 10: 10 -- -- -- -- -- -- -- 18 19 -- -- -- -- -- -- 20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 30: -- -- -- -- -- -- 36 -- -- -- -- -- -- -- -- -- 40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 4f 50: 50 51 -- -- -- -- -- -- -- -- -- -- -- -- -- -- 60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 70: -- -- -- -- -- -- -- 77
RAM SPD’s are start from address 0x50 and RAM temp sensors start from 0x18 at same bus. In my system, there are 2 DIMMs available. So address of 0x18 and 0x19 are DIMMs temp sensors.
After found this info, to read temperatures of RAM sticks, we need
jc42 kernel module loaded. After that you need to tell to module that which addresses are need to used. This process consists of writing
smbus_path. For example:
# modprobe jc42 # echo jc42 0x18 > /sys/bus/i2c/devices/i2c-0/new_device # echo jc42 0x19 > /sys/bus/i2c/devices/i2c-0/new_device
After that your ram sticks temperature will be visible on
jc42-i2c-0-19 Adapter: SMBus PIIX4 adapter port 0 at 0b00 temp1: +50.7°C (low = +0.0°C) ALARM (HIGH, CRIT) (high = +0.0°C, hyst = +0.0°C) (crit = +0.0°C, hyst = +0.0°C) jc42-i2c-0-18 Adapter: SMBus PIIX4 adapter port 0 at 0b00 temp1: +51.8°C (low = +0.0°C) ALARM (HIGH, CRIT) (high = +0.0°C, hyst = +0.0°C) (crit = +0.0°C, hyst = +0.0°C)
Reading SPD values from memory modules (optional)
To read the SPD timing values from memory modules, install the i2c-tools package. Once installed, load the
eeprom kernel module.
# modprobe eeprom
Finally, view memory information with
Here is partial output from one machine:
Memory Serial Presence Detect Decoder By Philip Edelbrock, Christian Zuckschwerdt, Burkart Lingner, Jean Delvare, Trent Piepho and others Decoding EEPROM: /sys/bus/i2c/drivers/eeprom/0-0050 Guessing DIMM is in bank 1 ---=== SPD EEPROM Information ===--- EEPROM CRC of bytes 0-116 OK (0x583F) # of bytes written to SDRAM EEPROM 176 Total number of bytes in EEPROM 512 Fundamental Memory type DDR3 SDRAM Module Type UDIMM ---=== Memory Characteristics ===--- Fine time base 2.500 ps Medium time base 0.125 ns Maximum module speed 1066MHz (PC3-8533) Size 2048 MB Banks x Rows x Columns x Bits 8 x 14 x 10 x 64 Ranks 2 SDRAM Device Width 8 bits tCL-tRCD-tRP-tRAS 7-7-7-33 Supported CAS Latencies (tCL) 8T, 7T, 6T, 5T ---=== Timing Parameters ===--- Minimum Write Recovery time (tWR) 15. 000 ns Minimum Row Active to Row Active Delay (tRRD) 7.500 ns Minimum Active to Auto-Refresh Delay (tRC) 49.500 ns Minimum Recovery Delay (tRFC) 110.000 ns Minimum Write to Read CMD Delay (tWTR) 7.500 ns Minimum Read to Pre-charge CMD Delay (tRTP) 7.500 ns Minimum Four Activate Window Delay (tFAW) 30.000 ns ---=== Optional Features ===--- Operable voltages 1.5V RZQ/6 supported? Yes RZQ/7 supported? Yes DLL-Off Mode supported? No Operating temperature range 0-85C Refresh Rate in extended temp range 1X Auto Self-Refresh? Yes On-Die Thermal Sensor readout? No Partial Array Self-Refresh? No Thermal Sensor Accuracy Not implemented SDRAM Device Type Standard Monolithic ---=== Physical Characteristics ===--- Module Height (mm) 15 Module Thickness (mm) 1 front, 1 back Module Width (mm) 133. 5 Module Reference Card B ---=== Manufacturer Data ===--- Module Manufacturer Invalid Manufacturing Location Code 0x02 Part Number OCZ3G1600LV2G ...
Using sensor data
There are a variety of front-ends for sensors data.
- psensor — GTK application for monitoring hardware sensors, including temperatures and fan speeds. Monitors motherboard and CPU (using lm-sensors), Nvidia GPUs (using XNVCtrl), and harddisks (using hddtemp or libatasmart).
- https://wpitchoune.net/psensor/ || psensor
- xsensors — X11 interface to lm_sensors.
- https://github.com/Mystro256/xsensors || xsensors
For specific Desktop environments:
- Freon (GNOME Shell extension) — Extension for displaying CPU temperature, disk temperature, video card temperature , voltage and fan RPM in GNOME Shell.
- https://github.com/UshakovVasilii/gnome-shell-extension-freon || gnome-shell-extension-freonAUR
- GNOME Sensors Applet — Applet for the GNOME Panel to display readings from hardware sensors, including CPU temperature, fan speeds and voltage readings.
- http://sensors-applet.sourceforge.net/ || sensors-applet
- lm-sensors (LXPanel plugin) — Monitor temperature/voltages/fan speeds in LXDE through lm-sensors.
- https://danamlund.dk/sensors_lxpanel_plugin/ || sensors-lxpanel-pluginAUR
- MATE Sensors Applet — Display readings from hardware sensors in your MATE panel.
- https://github.com/mate-desktop/mate-sensors-applet || mate-sensors-applet
- Sensors (Xfce4 panel plugin) — Hardware sensors plugin for the Xfce panel.
- https://goodies.xfce.org/projects/panel-plugins/xfce4-sensors-plugin || xfce4-sensors-plugin
- Thermal Monitor (Plasma 5 applet) — KDE Plasma applet for monitoring CPU, GPU and other available temperature sensors.
- https://gitlab.com/agurenko/plasma-applet-thermal-monitor || plasma5-applets-thermal-monitor
There is an optional daemon called sensord (included with the lm_sensors package) which can log data to a round robin database (rrd) and later visualize graphically. See the sensord(8) man page for details.
Tips and tricks
In some cases, the data displayed might be incorrect or users may wish to rename the output. Use cases include:
- Incorrect temperature values due to a wrong offset (i.e. temps are reported 20 °C higher than actual).
- Users wish to rename the output of some sensors.
- The cores might be displayed in an incorrect order.
All of the above (and more) can be adjusted by overriding the package provides settings in
/etc/sensors3.conf by creating
/etc/sensors.d/foo wherein any number of tweaks will override the default values. It is recommended to rename ‘foo’ to the motherboard brand and model but this naming nomenclature is optional.
Note: Do not edit
/etc/sensors3.conf directly since package updates will overwrite any changes thus losing them.
Example 1. Adjusting temperature offsets
This is a real example on a Zotac ION-ITX-A-U motherboard. The coretemp values are off by 20 °C (too high) and are adjusted down to Intel specs.
coretemp-isa-0000 Adapter: ISA adapter Core 0: +57.0°C (crit = +125.0°C) Core 1: +55.0°C (crit = +125.0°C) ...
sensors with the
-u switch to see what options are available for each physical chip (raw mode):
$ sensors -u
coretemp-isa-0000 Adapter: ISA adapter Core 0: temp2_input: 57.000 temp2_crit: 125.000 temp2_crit_alarm: 0.000 Core 1: temp3_input: 55.000 temp3_crit: 125.000 temp3_crit_alarm: 0. 000 ...
Create the following file overriding the default values:
chip "coretemp-isa-0000" label temp2 "Core 0" compute temp2 @-20,@-20 label temp3 "Core 1" compute temp3 @-20,@-20
sensors shows the adjust values:
coretemp-isa-0000 Adapter: ISA adapter Core 0: +37.0°C (crit = +105.0°C) Core 1: +35.0°C (crit = +105.0°C) ...
Example 2. Renaming labels
This is a real example on an Asus A7M266. The user wishes more verbose names for the temperature labels
as99127f-i2c-0-2d Adapter: SMBus Via Pro adapter at e800 ... temp1: +35.0°C (high = +0.0°C, hyst = -128.0°C) temp2: +47.5°C (high = +100.0°C, hyst = +75.0°C) ...
Create the following file to override the default values:
chip "as99127f-*" label temp1 "Mobo Temp" label temp2 "CPU0 Temp"
sensors shows the adjust values:
as99127f-i2c-0-2d Adapter: SMBus Via Pro adapter at e800 . .. Mobo Temp: +35.0°C (high = +0.0°C, hyst = -128.0°C) CPU0 Temp: +47.5°C (high = +100.0°C, hyst = +75.0°C) ...
Example 3. Renumbering cores for multi-CPU systems
This is a real example on an HP Z600 workstation with dual Xeons. The actual numbering of physical cores is incorrect: numbered 0, 1, 9, 10 which is repeated into the second CPU. Most users expect the core temperatures to report out in sequential order, i.e. 0,1,2,3,4,5,6,7.
coretemp-isa-0000 Adapter: ISA adapter Core 0: +65.0°C (high = +85.0°C, crit = +95.0°C) Core 1: +65.0°C (high = +85.0°C, crit = +95.0°C) Core 9: +66.0°C (high = +85.0°C, crit = +95.0°C) Core 10: +66.0°C (high = +85.0°C, crit = +95.0°C) coretemp-isa-0004 Adapter: ISA adapter Core 0: +54.0°C (high = +85.0°C, crit = +95.0°C) Core 1: +56.0°C (high = +85.0°C, crit = +95.0°C) Core 9: +60.0°C (high = +85.0°C, crit = +95.0°C) Core 10: +61.0°C (high = +85. 0°C, crit = +95.0°C) ...
sensors with the
-u switch to see what options are available for each physical chip:
$ sensors -u coretemp-isa-0000
coretemp-isa-0000 Adapter: ISA adapter Core 0: temp2_input: 61.000 temp2_max: 85.000 temp2_crit: 95.000 temp2_crit_alarm: 0.000 Core 1: temp3_input: 61.000 temp3_max: 85.000 temp3_crit: 95.000 temp3_crit_alarm: 0.000 Core 9: temp11_input: 62.000 temp11_max: 85.000 temp11_crit: 95.000 Core 10: temp12_input: 63.000 temp12_max: 85.000 temp12_crit: 95.000
$ sensors -u coretemp-isa-0004
coretemp-isa-0004 Adapter: ISA adapter Core 0: temp2_input: 53.000 temp2_max: 85.000 temp2_crit: 95.000 temp2_crit_alarm: 0.000 Core 1: temp3_input: 54.000 temp3_max: 85.000 temp3_crit: 95.000 temp3_crit_alarm: 0.000 Core 9: temp11_input: 59.000 temp11_max: 85.000 temp11_crit: 95.000 Core 10: temp12_input: 59.000 temp12_max: 85.000 temp12_crit: 95. 000 ...
Create the following file overriding the default values:
chip "coretemp-isa-0000" label temp2 "Core 0" label temp3 "Core 1" label temp11 "Core 2" label temp12 "Core 3" chip "coretemp-isa-0004" label temp2 "Core 4" label temp3 "Core 5" label temp11 "Core 6" label temp12 "Core 7"
sensors shows the adjust values:
coretemp-isa-0000 Adapter: ISA adapter Core0: +64.0°C (high = +85.0°C, crit = +95.0°C) Core1: +63.0°C (high = +85.0°C, crit = +95.0°C) Core2: +65.0°C (high = +85.0°C, crit = +95.0°C) Core3: +66.0°C (high = +85.0°C, crit = +95.0°C) coretemp-isa-0004 Adapter: ISA adapter Core4: +53.0°C (high = +85.0°C, crit = +95.0°C) Core5: +54.0°C (high = +85.0°C, crit = +95.0°C) Core6: +59.0°C (high = +85.0°C, crit = +95.0°C) Core7: +60.0°C (high = +85.0°C, crit = +95.0°C) ...
Automatic lm_sensors deployment
Users wishing to deploy lm_sensors on multiple machines can use the following to accept the defaults to all questions:
# sensors-detect --auto
M.A.R.T. drive temperature
Since kernel 5.6 the
drivetemp module will report SATA/SAS temperature through hwmon, but
sensors-detect does not automatically detect this so the module must be manually loaded.
# modprobe drivetemp
You should now see entries similar to this in your
drivetemp-scsi-1-0 Adapter: SCSI adapter temp1: +33.0°C drivetemp-scsi-2-0 Adapter: SCSI adapter temp1: +32.0°C (low = +0.0°C, high = +70.0°C) (crit low = +0.0°C, crit = +70.0°C) (lowest = +29.0°C, highest = +41.0°C)
Configure automatic module loading to load the module on boot.
Some K10 processors have issues with their temperature sensor. See the k10temp documentation for more information.
On affected machines the module will report «unreliable CPU thermal sensor; monitoring disabled». To force monitoring anyway, you can run the following:
# rmmod k10temp # modprobe k10temp force=1
Confirm that the sensor is in fact valid and reliable. If it is, can edit
/etc/modprobe.d/k10temp.conf and add:
options k10temp force=1
This will allow the module to load at boot.
Asus B450M-A/A320M-K/A320M-K-BR motherboards
These motherboards use a IT8655E chip, which is not supported by the it87 kernel driver, as of Nov 2020 . However, it is supported by the upstream version of the kernel driver . The DKMS variant is contained in it87-dkms-gitAUR.
Asus B450/X399/X470 motherboards with AM4 Socket
Some recent Asus motherboards use a ITE IT8665E chip, accessing the temperature, fan and voltage sensors may require the
asus-wmi-sensors module. It is part of the mainline kernel since 5.17: load the
asus-wmi-sensors kernel module which uses the UEFI interface and may require a BIOS update on some boards .
it87 module reads the values from the chip directly, install it87-dkms-gitAUR and load the
it87 kernel module.
Asus H97/Z97/Z170/Z370i/X570/B550 motherboards
With some recent Asus motherboards, fan and voltage sensor access may require the
nct6775 kernel module to be loaded.
You may also need to add the following kernel parameter:
See https://bugzilla.kernel.org/show_bug.cgi?id=204807 for more information.
Note: Starting with Kernel 5.16 , the above kernel parameter is no longer be required for most boards and should be avoided.
Asrock Deskmini h570
The STX board of the Deskmini h570 uses a NCT6683 chip, for accessing the temperature, fan and voltage sensors the loading of
nct6683 module is required.
For proper values of the
nct6683 module have a module config file created:
options nct6683 force=1
Gigabyte B250/Z370/B450M/B560M/Z690 motherboards
Some Gigabyte motherboards use the ITE IT8686E, ITE8689 (for B560) or ITE8689E (for Z690) chip, which is not supported by the it87 kernel driver, as of May 2019 . However, it is supported by the upstream version of the kernel driver . The DKMS variant is contained in it87-dkms-gitAUR. As with #Asus H97/Z97/Z170/Z370i/X570/B550 motherboards, a kernel parameter is required before attempting to install the module:
Furthermore, supply the id of the chip when loading the module as follows:
# modprobe it87 force_id=0x8686 or # modprobe it87 force_id=0x8689 # for B560 # modprobe it87 force_id=0x8628 # for Z690
Or you can load the module during boot process by creating the following two files:
options it87 force_id=0x8628
options it87 ignore_resource_conflict=1
Once the module is loaded you can use the sensors tool to probe the chip.
Now you can also use fancontrol to control the speed step of your case fan.
Optionally installation of zenpower3-dkmsAUR may allow greater fine tuning of the motherboard’s cooling system. However, it does disable the default k10temp module.
This motherboard uses the ITE IT8620E chip (useful also to read voltages, mainboard temp, fan speed). As of October 2014, lm_sensors has no driver support for chip ITE IT8620E  . lm_sensors developers had a report that the chip is somewhat compatible with the IT8728F for the hardware monitoring part. However, as of August 2016,  lists the IT8620E as supported.
You can load the module at runtime with modprobe:
$ modprobe it87 force_id=0x8728
Or you can load the modules during boot process by creating the following two files:
options it87 force_id=0x8603
Once the module is loaded you can use the sensors tool to probe the chip.
Now you can also use fancontrol to control the speedsteps of your case fan.
Laptop screen issues after running sensors-detect
This is caused by lm-sensors messing with the Vcom values of the screen while probing for sensors. It has been discussed and solved at the forums already: https://bbs.archlinux.org/viewtopic.php?id=193048. However, make sure to read through the thread carefully before running any of the suggested commands.
i2c bus errors on AMD Navi 2 GPUs
There is currently a bug in the way the kernel handles reading the i2c bus on AMD Navi 2 GPUs. The bus currently can only be used with EEPROMs and trying to use it with other devices will cause it to fail. This can cause crashes, black screens, and even cause the card to behave oddly like unable to switch power states. Its currently advised not to scan the i2c bus if you have a Navi 2 based card. You can read more here: https://gitlab.freedesktop.org/drm/amd/-/issues/1470
Optimal CPU & GPU Temperature [PC & Laptops]
Looking for a list of Optimal GPU and CPU Temperature?
Then you’ve arrived at the right place. There can be many factors that can cause a Computer to overheat.
If you’re worried that your PC Temperature is too high, this guide will help you out to find whether it is optimal or not. And, this article applies to both the Desktops and Laptops.
I’ve mentioned the Average Temperate Range of both CPU and GPU and the safe temperature limit as well. And in case your system’s temperature is close to the max recommended value, there are various tips here that you can apply to lower down the temperature to a great extent.
How Hot Should my CPU be?
Well, the answer is not as simple as you think. Every Processor is different and depending on the ambient room temperature and the cooling solution used, the CPU Temperature may vary.
During General Usage, your CPU Temperature can have any value between 30-65 °C.
But with Extended Gaming and Heavy Usage, the temperature can easily reach a range of 65-90 °C.
How Hot is Too Hot for Your CPU?
Every Processor has a maximum safe temperature limit. I’ve listed the maximum temperature of almost all the recent Processors of Intel and AMD. You can scroll down and look for your Processor on the list.
But for most Processors, the maximum temperature limit is somewhere around 90-95 °C.
Here’s a video by Linus to help you understand Safe PC Temperatures better.
As you are now aware of the problems that can be caused due to high temperature, let’s dig into whether your CPU Temperature is too high or not.
First up, we have the Ideal & Max Temperature Range for CPUs. Later on, we will look at the optimal temperature for Graphics Cards and HDD/SSD.
- 1 Optimal CPU Temperature
- 2 Maximum CPU Temperature
- 3 GPU Temperature
- 4 Optimal Hard Disk / SSD Temperature
- 5 How To Monitor CPU & GPU Temperature
- 6 Cooling Solutions for PC and Laptop
- 7 References
Optimal CPU Temperature
In General Usage, most processors will have an average temperature between 40-70 °C.
The lower the temperature, the better. Low temperature is recommended for the efficient performance of your Computer.
Here is the optimal temperature range for the most popular Processor Series of Intel and AMD.
|Processor Series||Average Temperature Range|
|Intel Core i7||50-65 °C|
|Intel Core i5||50-62 °C|
|Intel Core i3||50-60 °C|
|Intel Core 2 Duo||45-55 °C|
|Intel Pentium Pro||75-85 °C|
|Intel Pentium Mobile||70-85 °C|
|Intel Pentium 4||45-65 °C|
|Intel Pentium 3||60-85 °C|
|Intel Celeron||67-85 °C|
|AMD A10||50-60 °C|
|AMD A6||45-57 °C|
|AMD Athlon||85-95 °C|
|AMD Athlon 64||45-60 °C|
|AMD Athlon FX||45-60 °C|
Credits: Computer Hope
The above table can give a basic idea of the Normal CPU Temperature in most scenarios.
Note: This is the general range of the temperature for most processors. However, the actual Temperature may vary from model to model and from generation to generation. Also, the ambient temperature and the cooling solution being used will affect the average temperature.
If your Processor’s Temperature falls in this range (or lower than it), there’s absolutely no reason to worry.
But if the CPU Temperature is higher, you should check the next section and ensure that it is lower than the maximum temperature stated by the manufacturer. Anything lower than the Maximum Temperature can be considered safe.
Maximum CPU Temperature
Both the Intel and AMD have a specified maximum temperature value which is mentioned in your Processor’s Specification. The same value is also mentioned down below.
During heavy usage, it is a must to ensure that the CPU Temperature stays below this number.
For Intel Processors
Here is a list of Intel 8th, 7th, 6th, and 5th Generation Desktop & Laptop Processors along with their maximum temperature.
If you’re not sure about your Processor’s Model Number, Download CPU-Z and it will display all the details of your Processor.
According to Intel, Junction Temperature (TJUNCTION) is the maximum temperature allowed at the processor die and Case Temperature (Tᴄᴀsᴇ) is the maximum temperature allowed at the processor Integrated Heat Spreader.
In the following table, all the values are for TJUNCTION temperature. But for a few processors, I’ve used Tᴄᴀsᴇ and have mentioned it there if that particular value is for Case Temperature.
|Series||Processor Name||Maximum Temperature|
|10th Generation||Core i9
|Core i7||10710U, 1065G7, 1060G7||100°C|
|Core i5||10510U, 10210U, 1035G7, 1035G4, 1035G1, 1030G7, 1030G4||100°C|
||10110U, 1005G1, 1000G4, 1000G1||100°C|
|9th Generation||Core i9||9920X, 9900X||92°C|
|9900KS, 9900K, 9900KF, 9000, 9880H, 9980HK||100°C|
|9700K, 9700KF, 9700F, 9700, 9750HF, 9750H, 9850H||100°C|
|Core i5||9600K, 9600KF, 9400F, 9400, 9300H, 9400H||100°C|
|Core i3||9350KF, 9300, 9100, 9100F||100°C|
|Pentium Gold||G5620, G5420||100°C|
|8th Generation||Core i9||8950HK||100°C|
|Core i7||8086K, 8700K, 8700, 8700T. 8850H, 8750H 8559U, 8650U, 8550U, 8809G. 8709G, 8706G, 8705G, 8305G||100°C|
|Core i5||8600K, 8600, 8600T, 8500, 8500T, 8400, 8400T, 8400H, 8300H 8269U, 8259U, 8350U, 8250U||100°C|
|8350K, 8300, 8100, 8109U, 8130U||100°C|
|Pentium Gold||G5600, G5500, G5400||100°C|
|7th Generation||Core i9||7940X||102°C|
|Core i7||7800X, 7740X, 7700K, 7700, 7920HQ, 7820HQ, 7820HK & 7700HQ, 7660U, 7600U, 7567U, 7560U, 7500U, i7-7Y75||100°C|
|Core i5||7640X, 7600K, 7600, 7500, 7400 & 7440HQ, 7300HQ, 7360U, 7300U, 7287U, 7267U, 7260U, 7200U, 7Y57, 7Y54||100°C|
|7600T, 7500T, 7400T||80°C|
|Core i3||7350K, 7320, 7300, 7100, 7101E, 7100H & 7167U, 7130U, 7100U||100°C|
|Core m3||7Y32, 7Y30||100°C|
|Pentium||G4620, G4600, G4560, 4415U, 4410Y||100°C|
|Pentium Silver||J5005, N5000||105°C|
|Celeron||G3950, G3930, 3965U, 3865U||100°C|
|J4105, J4005, N4100, N4000||105°C|
|6th Generation||Core i7||6700K,||Tᴄᴀsᴇ = 64°C|
|6785R, 6700||Tᴄᴀsᴇ = 71°C|
|6700T||Tᴄᴀsᴇ = 66°C|
|6970HQ, 6920HQ, 6870HQ, 6820HQ,
6770HQ, 6700HQ, 6660U, 6650U,
6600U, 6567U, 6560U, 6500U
||6600K||Tᴄᴀsᴇ = 64°C|
|6685R, 6600, 6585R, 6500, 6402P, 6400||Tᴄᴀsᴇ = 71°C|
|6600T, 6500T, 6400T||Tᴄᴀsᴇ = 66°C|
|6440HQ, 6360U, 6350HQ, 6300HQ,
6300U, 6287U, 6267U, 6260U, 6200U
||6320, 6300, 6100||Tᴄᴀsᴇ = 65°C|
|6300T, 6100T, 6098P,||Tᴄᴀsᴇ = 66°C|
|6167U, 6157U, 6100H, 6100U, 6006U||100°C|
|Core m5||6Y57, 6Y54||100°C|
|Celeron||G3902E, G3900E, 3955U, 3855U||100°C|
|5th Generation||Core i7||5950HQ, 5850HQ, 5750HQ, 5700HQ,
5650U, 5600U, 5557U, 5550U, 5500U
|Core i5||5350H, 5350U, 5300U, 5287U, 5257U, 5250U, 5200U||105°C|
|Core i3||5157U, 5020U, 5015U, 5010U, 5005U||105°C|
|Core M||5Y71, 5Y70, 5Y51, 5Y31, 5Y10c, 5Y10a, 5Y10||95°C|
To know the Maximum Temperature of Previous Generations, visit Intel’s Official Website.
As you can see, in case of most of the Processors, the maximum temperature is 100°C. Just for some extra precaution, ensure that your CPU Temperature is always at least 10°C lower than the maximum value.
For AMD Processors
Here’s the list for AMD’s CPUs and APUs. AMD hasn’t mentioned the temperature limit for all their processors but I managed to find it for the latest generations and that includes Ryzen.
|Series||Processor Name||Maximum Temperature|
|Ryzen 4000 Series (Zen 3 and Zen 2)||Ryzen 7
||4800H, 4800U, 4700U||105°C|
|Ryzen 5||4600H, 4600U, 4500U||105°C|
|Ryzen 3000 Series (Zen 2 and Zen+)||Threadripper||3990X, 3970X, 3960X||95°C|
|Ryzen 7||3700X, 3800X||95°C|
||3400G, 3600, 3600X||95°C|
|Ryzen 2000 Series (Zen+ and Zen)||Threadripper||2920X, 2950X, 2970WX, 2990WX||68°C|
|2700U, Pro 2700U, 3750H||105°C|
|Ryzen 5||2600, 2600X||95°C|
|2400GE, Pro 2400GE, 2400G, Pro 2400G, 2500U, Pro 2500U||105°C|
|Ryzen 3||2200GE, Pro 2200GE, 2200G, Pro 2200G, 2300U, Pro 2300U, 2200U||105°C|
|Ryzen 1000 Series (Zen)||Threadripper||1950x, 1920x, 1900x||68°C|
|Ryzen 7||1800x, Pro 1700x, Pro 1700, 1700||95°C|
|Ryzen 5||1600X, Pro 1600, 1600, 1500X, Pro 1500, 1400||95°C|
|Ryzen 3||1300X, Pro 1300, Pro 1200, 1200||95°C|
|Bristol Ridge||A12||9800, 9800E, Pro 9800, Pro 9800E||90°C|
|A10||9700, 9700E, Pro 9700E, Pro 9700||90°C|
|A8||9600, Pro 9600||90°C|
|A6||9550, 9500, 9500E, Pro 9500, Pro 9500E||90°C|
For the Ryzen Processors, there’s something important that you need to keep in mind. In Ryzen Master Utility, the Temperature that is being reported is around 20°C higher than the actual junction temperature. Here’s a news from AMD that confirms the same.
So, if in the Ryzen Master Utility the reported temperature is 65°C, then the actual junction temperature is just 45°C.
The Temperature of the GPU usually stays in a normal range unless you start a GPU intensive job.
Optimal Temperature of an idle GPU should be around 35-55 °C. During Gaming and Heavy GPU Usage, the average GPU Temperature is generally around 60-80°C.
A major problem with the GPU Temperature can arise when you play games for an extended period of time. It is necessary to ensure that even after a long period of gaming, the GPU Temperature stays at least 5-10°C lower than the maximum temperature specified by the OEM.
For Nvidia GPU
Here is the Maximum GPU Temperature for Nvidia GeForce 10, 900, 700, and 600 Series.
For older Graphics Cards, you can look for it on Nvidia’s official website. If you’re unable to find it on their website, just comment below and I’ll look it up for you.
|Series||GPU Name||Maximum GPU Temperature|
|GeForce 20 Series||Titan RTX||89°C|
|RTX 2080 Ti||89°C|
|RTX 2080 Super||89°C|
|RTX 2070 Super||88°C|
|RTX 2060 Super||89°C|
|GeForce 16 Series||GTX 1660 Ti||95°C|
|GeForce 10 Series||TITAN V||91°C|
|GTX 1080 Ti||91°C|
|GTX 1070 Ti||94°C|
|GTX 1050 Ti||97°C|
|GeForce 900||TITAN X||91°C|
|GTX 980 Ti||92°C|
|GeForce 700||GTX TITAN Z||95°C|
|GTX TITAN Black||95°C|
|GTX 780 Ti||95°C|
|GTX 750 Ti||95°C|
|GeForce 600||GTX 690||98°C|
|GTX 660 Ti||97°C|
|GTX 650 Ti||105°C|
On average, the Maximum Temperature limit for Nvidia’s Graphics Cards is between 90°C and 100°C. The GTX 650 Ti and GT 610 are the exceptions for this rule as their max temps is 105°C and 102°C respectively.
For most systems, try to ensure that the GPU temperature doesn’t go over 85°C.
For AMD GPU
Even though AMD hasn’t specified a maximum temperature limit for their Graphics Cards on their website, I confirmed from various sources that Temperature limit of AMD Radeon GPUs is around 90-95°C and that is same as the most Nvidia GPUs.
|AMD Radeon Series||90-95°C|
Once again, make use of sufficient cooling to keep the temperature below 85°C for safe working.
Optimal Hard Disk / SSD Temperature
Should you bother about the Temperature of your Hard Disk or Solid State Drive?
Well, in most cases, the answer is no. It is very less likely that your HDD/SSD Temperature will go anywhere near the maximum value.
But still in many cases, it could be possible and it severely affects the performance of your Disk. In the long run, it can also lead to disk failure.
Here’s the Operating Temperature limit for most popular HDDs and SSDs. If the one you own is not present in this list, you can find the details in the user manual of your product.
|WD Blue HDD||0-60°C|
|Seagate Barracuda HDD||0-60°C|
|ADATA Ultimate SU800 SSD||0-70°C|
|Samsung Evo 860 SSD||0-70°C|
|WD Green SSD||0-70°C|
|Samsung Evo 960 m.2 SSD||0-70°C|
|Intel Optane Memory||0-70°C|
In case of most Internal Storage Drives, the temperature range is usually between 0°C to 60°C or from 0°C to 70°C.
How To Monitor CPU & GPU Temperature
There are various tools like AIDA64, Speccy, and HWMonitor by using which you can easily monitor your PC Temperature.
I personally prefer HWMonitor and it is free as well.
You can easily monitor the CPU, GPU, and Hard Disk Air Flow Temperature using this tool. It also monitors the maximum and minimum Temperature that the component has reached since you started HW Monitor. If you want to reset the Min/Max Stats, go to View in Menu and select Clear Max/Min.
Here’s a Screenshot of HWMonitor.
Monitor Temperature While Gaming
The HWMonitor is a great tool but you can’t use it in-game.
Here’s when the MSI Afterburner comes to the rescue. It not only displays the Utilization and Temperature of your Components, but it can also show the in-game FPS.
After Installing MSI Afterburner, open it and click on the Gear (Settings icon).
This will take you to the MSI Afterburner Properties.
Inside the Monitoring Tab, click on the Components that you want to be displayed during the game.
Now tick the “Show in On-Screen Display” checkbox for every component that you want to be displayed.
When you’re done press OK. Make sure that MSI Afterburner is still open.
Now go to Windows Taskbar and click on the RivaTuner Statistics Server. This should look like the one shown in the image below.
Make sure that On-Screen Display support is On. By using the On-Screen Display Palette, you can change the color of the text that appears inside the game.
Now, move around the On-Screen Display Zoom slider depending on how big or small you want the text to appear in the game.
Once, you’re done setting it up. Start your game to monitor the thermals.
What to do if my CPU/GPU Temperature is not within the safe limits?
There could be many reasons behind an abnormal temperature value. Before we can come to any conclusions, move to the next section and apply the PC Cooling Tips. If you don’t succeed, it is better to visit a Computer Store or Service Center to get your PC checked.
Cooling Solutions for PC and Laptop
We hope you’re finding our content helpful. We just wanted to let you know that some links on this page are Amazon Affiliate Links and we may earn a small commission when you purchase something after visiting these links. To learn more, please visit this page.
The following Cooling Tips can significantly lower the temperature of your PC or Laptop. Make sure to try as many as possible and each one of them will contribute to the thermal levels.
1) Clean Dust
Remove all the dust inside your CPU Case by using an Air Blower. Do it slowly ensuring that you don’t damage any components. If you use a Laptop, use Compressed Gas to gid rid of Dust. But the more efficient way is to open the Laptop and then remove the dust.
Make sure that the Case Fan as well as the CPU Fan is free from dust and is working optimally.
2) Apply a New Thermal Paste
Use Isopropyl Alcohol to clean the existing Thermal Paste and replace it with a new one. The Arctic Silver is one of the most trusted and best Thermal Compounds out there. Don’t apply too much Thermal Paste either. A pea-sized amount is the most effective technique to apply thermal paste.
3) Upgrade to a Better CPU Cooler
A good fan can keep the CPU Temperature significantly lower. If you’re using a stock cooler, it is better to get rid of it and use a better one. The Cooler Master Hyper RR-212E-20PK-R2 will do an excellent job. It is one of the most effective CPU Coolers out there.
4) Add More Case Fans
If your Case has lower than 3 fans, add a few more for more optimal air-flow to and from the components. The CPU Cooler only works on the CPU while the Case Fans will take care of all the PC Components.
5) Get a Laptop Cooling Pad
So, we already had enough talk on PC Fans and Coolers. So, what about Laptops?
The Laptops Cooling Pads work very well in keeping the CPU Temperature low. We actually have a guide on Laptop Cooling Pads and Tips.
6) Make Your Room Cool
One of the biggest cause of PC Overheating is the Surrounding Temperature. You may not realize it but the temperature difference of PC Components during Summers and Winters can be huge.
Turn On your Room Cooler or Air Conditioner and it will reduce the CPU Temperature.
NUC CPU temperature concerns — Hardware
I’m running HA on an Intel NUC i7 (about 6 months old) and I notice the fan often runs and CPU temp is about 74⁰C at idle, often going up over 85⁰C when I so something like restart HA. The NUC is sitting on top of a cabinet in a cool room (currently the room is probably only about 15⁰C) so I’m worried that this CPU temp is far too high for the low load.
CPU usage is currently about 18%.
I’m wondering what other people are seeing as NUC CPU temps and what I should look at to reduce this. First thoughts are checking the heatsink paste which I will try to do this week.
This does sound like a little hot for NUC. Even for an i7. I suppose given the 6 months old that it is a gen8? What are you running on it? If it is HA on ubuntu then it is way too hot. I had a gen 6 idling at 46C under windows. Maybe look at removing the cooler and changing out the thermal paste?
I have the gen 8 i7 (nuc8i7beh) and have a current idle temperature of 37. Running ubuntu and HA under docker together with 13 other containers.
Try update your bios if you haven’t allready and check your fan settings.
Check setting in the bios as well, speedstep I think it was called.
In fact sit in the bios for a bit and see what it reports the cpu temperature as.
Thanks for the responses everyone. It is an 8th gen running Ubuntu, docker / HA (old hassio) with Motioneye add-on and a bunch of other add-ons like Google drive backup, gmusic, mqtt, assistant relay, ESPhome, VSC, Unifi controller, Node-red & Portainer (and a couple of others).
I’ll open up the NUC today as I bought some heatsink paste yesterday and will check bios too
Update: I replaced the heat sink paste and updated the BIOS firmware but the CPU temp still hovers in the same +80C range, basically at idle. While I was in the BIOS menu doing the update the CPU was only at around 40C but obviously this was with almost no load at all.
Currently the NUC is showing the below via SSH:
~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +80.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +57.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +53.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +80.0°C (high = +100.0°C, crit = +100. 0°C) Core 3: +75.0°C (high = +100.0°C, crit = +100.0°C)
What temp settings should I have for the fan control? currently they are just default but I’m wondering if these are no good… going to try it on ‘Cool’ mode or even drop the min. temp further yet. But I’m confused as to why it runs so hot in the first place. It was spotlessly clean when I opened it up, no dust at all
Because you have 2 cores very high and two at relatively reasonable temperature, I suspect that either your cooler is not sitting correctly on the CPU or that you have a defective cpu.
Do you consistently see those same 2 cores being high? Was this the case before you took out the cooler and replaced the TIM?
Its always run hot since I got it. I was very careful to put the cooler back properly when I replaced the paste this morning. I also changed the fan settings to try and bring the temp down, which it has a bit, but it’s still up around 65~85C all the time. Using the sensors command in SSH I see widely varying temps. I can refresh it within seconds and see temps 20C different…
Not sure these figures can be believed… (all taken within seconds)
[email protected]:~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +65.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +63.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +54.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +67.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +57.0°C (high = +100.0°C, crit = +100.0°C) [email protected]:~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +100.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +64.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +57. 0°C (high = +100.0°C, crit = +100.0°C) Core 2: +97.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +62.0°C (high = +100.0°C, crit = +100.0°C) [email protected]:~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +84.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +57.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +84.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +68.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +58.0°C (high = +100.0°C, crit = +100.0°C) [email protected]:~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +70.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +68.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +67.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +70.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +68.0°C (high = +100.0°C, crit = +100.0°C) [email protected]:~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +92. 0°C (high = +100.0°C, crit = +100.0°C) Core 0: +92.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +62.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +64.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +64.0°C (high = +100.0°C, crit = +100.0°C) [email protected]:~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +67.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +67.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +56.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +61.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +57.0°C (high = +100.0°C, crit = +100.0°C) [email protected]:~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +59.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +59.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +57.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +58.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +57.0°C (high = +100. 0°C, crit = +100.0°C) [email protected]:~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +87.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +63.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +60.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +63.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +87.0°C (high = +100.0°C, crit = +100.0°C) [email protected]:~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +87.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +63.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +60.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +63.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +87.0°C (high = +100.0°C, crit = +100.0°C) [email protected]:~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +77.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +67.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +64.0°C (high = +100. 0°C, crit = +100.0°C) Core 2: +75.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +66.0°C (high = +100.0°C, crit = +100.0°C) [email protected]:~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +57.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +57.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +55.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +54.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +53.0°C (high = +100.0°C, crit = +100.0°C) [email protected]:~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +93.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +93.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +71.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +92.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +66.0°C (high = +100.0°C, crit = +100.0°C) [email protected]:~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +55.0°C (high = +100. 0°C, crit = +100.0°C) Core 0: +55.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +53.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +53.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +53.0°C (high = +100.0°C, crit = +100.0°C) [email protected]:~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +73.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +73.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +56.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +73.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +70.0°C (high = +100.0°C, crit = +100.0°C) [email protected]:~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +82.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +82.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +56.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +69.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +66.0°C (high = +100.0°C, crit = +100. 0°C) [email protected]:~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +65.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +65.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +61.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +49.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +50.0°C (high = +100.0°C, crit = +100.0°C) [email protected]:~$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +74.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +52.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +51.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +74.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +62.0°C (high = +100.0°C, crit = +100.0°C)
Show us the ”top” command as well.
The hot cores seem to be moving around so it is either a gross physical problem or as @Cerb is suggesting, you have some rogue programs loading your CPU in pulses.
I had problem with NUC8i7 getting hot to 100C and fan was going max. The fix was: disable turboboost
To disable it, Enter the BIOS setup and from the System Utilities screen, select System Configuration. Then, navigate to BIOS /Platform Configuration (RBSU) > Performance Options > Intel ® Turbo Boost Technology and press Enter.
I was curious to my CPU temp (never bothered before) so I checked it too. I’m on an older NUC3 with an Intel N3700 CPU. Running Ubuntu 18.04, HA, Supervisor, Motioneye (1x 720p can for now), unifi, adguard, vscode, influxdb, mariadb, node-red and deconz.
[email protected]:~$ sensors acpitz-virtual-0 Adapter: Virtual device temp1: +58.0°C (crit = +115.0°C) coretemp-isa-0000 Adapter: ISA adapter Core 0: +55.0°C (high = +90.0°C, crit = +90.0°C) Core 1: +49.0°C (high = +90.0°C, crit = +90.0°C) Core 2: +47. 0°C (high = +90.0°C, crit = +90.0°C) Core 3: +47.0°C (high = +90.0°C, crit = +90.0°C)
This is pretty constant even when I view the camera in Motioneye.
Intel quite a while ago change the TIM (thermal interface material) they use on the lid of the CPUs. They now use cheaper paste, apart from I think the big boy chips (I think).
But those two cores look rather hot, not sure what else you could do tho.
Did you tighten the heatsink screws the same amount ? Uneven pressure would cause a temperature difference.
Sorry but this is not what I would call a fix. It’s a recommendation I have seen from… Dell laptop users because their cooling is so abysmally poor that they have to disable performance features on their CPU to prevent them from throttling or overheating. This is absurd. Buy a $500 system to make it run like a $200 one because they saved $2 on their cooling engineering. (I have one and posted one such experience their forum too.)
System run normal after disabling, no cpu power loss. Turbo boost is creating big peaks on cpu load, temeperature to 100c and noisy fan, now I have a smooth running system. I am running debian os and zoneminder in docker.
No, you definitely have performance loss. Turbo Boost is designed to boost the CPU clock for urgent short term needs. The CPU has a built in period of time during which that boost is maintained and is normally calibrated in the BIOS. It is designed to do this with normal cooling and improve responsiveness and a very legitimate performance feature. Disabling it is like disabling a core out of four and say that because you don’t need the 4th core in a single threaded benchmark, you have no performance problem. It is abnormally poor cooling which is causing you to have temperature spikes and noisy fan. I never disable this on any of my PCs or NUCs. Never had temperature go above 60C under load with factory cooling. Except for the stupid Dell laptop for which I had to modify the cooling and managed to get the temperature down by 35C at idle. Turbo boost is on and under full load of 5min I never hear the fan.
I’ll look at the turbo boost to see what happens if nothing else. What I don’t understand is the temps being so high with CPU usage less than 20%
This is zoneminder livestream with turbo enabled, max power and quiet fan setting, but fan is going loud:
$ sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +100.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +88.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +84.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +100.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +84.0°C (high = +100.0°C, crit = +100.0°C) acpitz-acpi-0 Adapter: ACPI interface temp1: -263.2°C temp2: +27.8°C (crit = +119.0°C) iwlwifi-virtual-0 Adapter: Virtual device temp1: N/A pch_cannonlake-virtual-0 Adapter: Virtual device temp1: +77. 0°C
And this livestream without turbo, max power, quiet fan settings, nice quiet without to me noticeable cpu power loss:
sensors coretemp-isa-0000 Adapter: ISA adapter Package id 0: +75.0°C (high = +100.0°C, crit = +100.0°C) Core 0: +69.0°C (high = +100.0°C, crit = +100.0°C) Core 1: +75.0°C (high = +100.0°C, crit = +100.0°C) Core 2: +71.0°C (high = +100.0°C, crit = +100.0°C) Core 3: +67.0°C (high = +100.0°C, crit = +100.0°C) acpitz-acpi-0 Adapter: ACPI interface temp1: -263.2°C temp2: +27.8°C (crit = +119.0°C) iwlwifi-virtual-0 Adapter: Virtual device temp1: N/A pch_cannonlake-virtual-0 Adapter: Virtual device temp1: +68.0°C
Both datasets are insanely hot. You have a cooling hardware problem. Again, what you are doing is equivalent of this: “My tooth hurts!- I have a fix: Pull them all out!”. Of course the hurting stops but you can’t use them to eat anymore.
My NUC runs below 45C idle and 75C under load (using stress) with ambient at 23C with everything enabled.
You do not run idle zoneminder with 4 cameras up to 2K with modect
next page →
- Small digression
- What are the ways to measure temperature?
- Thermal Monitor — from comparator to TM1/2
- Core 2 vs K8
The text will use the term ‘measurement’, but everyone knows that measuring something requires calibrated and verified instruments, and finding such at home is somewhat difficult. For non-calibrated instruments (accuracy 5 and above), the term ‘indicator’ should be used… but not everyone will immediately understand what it is about. Programmatic ways do not have accuracy a priori.
- finger on the radiator
- by smell and smoke
- software from the temperature sensor on the motherboard
- software from the thermal diode on the motherboard
- programmatically from a diode in a
- programmatically from the internal processor temperature meter (DTS)
Well, we will omit the second option, although it has the right to life — sometimes components with a low melting point are specially used. A typical example is a fire alarm sensor.
The first option, despite its seeming stupidity, gives quite good accuracy and reliability. In fact, any person can easily identify a temperature of about 36 degrees. 40-45 degrees feels ‘hot’, 50-55 feels ‘very hot’, and 60 degrees and above just can’t hold a finger. Interestingly, 30-60 degrees just corresponds to the normal temperature range of a processor heatsink.
The third option seems to be reliable, but it is precisely this that can give even worse reliability than the first method. The fact is that an element with a non-linear dependence of its characteristics on temperature is used as a measuring sensor. Actually, it is not difficult to very accurately convert resistance into temperature, but no one does this. In monitoring microcircuits, the principle of tabular approximation is used, in which the temperature is assigned to the resistance of the thermistor (more precisely, to the input voltage). The table is not for all values, much smaller, a linear approximation is made between the table values. Simple, clear and completely inaccurate. Because there is a conjugation of linear and non-linear parameters, then this method can give the effect of skipping and doubling the values. More precisely, not ‘can’, but gives. With a monotonous increase in temperature, quite often I observed slipping of some numbers, something like 45-45. 5-46-47-47.5. Probably, there were sequences -47.5-47.5-, but it is difficult to distinguish them ‘at a glance’ from 47-47.5-48. To the conversion errors, we must add the spread of parameters and the temporal instability of the thermistor. Monitoring is configured for some type of sensor, but during the production of the motherboard, the thermistor can be replaced with a similar one from other manufacturers, which will additionally introduce distortions into the measured values. How many times has it been that when a new BIOS version was released, the temperature readings moved down.
The fourth option is free from the disadvantage of the third, the sensor on a diode (p-n junction) in the temperature range of 0 … 70 (up to 100, depending on the package) degrees has linear characteristics. With constant current, the change in voltage across it is ‘-2mV’ times the temperature. The number ‘-2mV’ is a constant for a silicon semiconductor, the sign ‘-‘ indicates that when heated, the voltage at the p-n junction decreases. Converting voltage to temperature is not difficult, just subtract some offset Vo and multiply by the conversion factor. Subtraction and multiplication is done on a conventional amplifier. That. there is a temperature measurement tool that has extremely high accuracy and repeatability (compared to a thermistor), but, alas, not without a drawback — when measuring temperature, Vo must be subtracted, but it depends very much on many parameters! The most significant are the dimensions of the p-n junction and the current through it. The more powerful the semiconductor element, the less voltage drops across it at the same current. This means that different semiconductor elements can be used in the same circuit, it is only important to choose the current to maintain the same Vo. Usually, Vo is set in the monitoring chip, and when replacing the sensor, it is better not to change this offset (although it can be adjusted to a small extent). Alas, this method of measuring temperatures in the RS is not used. Although most monitoring chips can work with resistive and diode sensors, the latter are not used. Personally, it is difficult for me to understand why a thermistor is preferable to a diode or any transistor in a diode connection, except for one thing — the conservatism of the developers.
The fifth option differs from the fourth one only in that the sensor is located directly on the processor chip (I have not seen a variant with the temperature sensor placed in the processor case, but not on the chip). The technology is quite well developed and the repeatability of the parameters of semiconductor structures on a chip is very high. This can guarantee the same parameters of sensors for all processors of the same family, and with due attention of processor developers, for different families of the same company. If sclerosis does not change 🙂 , Pentium2/3 processors got the first sensors on the PC. At first, AMD had a hard time with this, the diode appeared only in the Palomino model. This diode served a slightly different purpose, its task was to limit the temperature of the processor — turn off the power when overheated. It seems to be a trifle, but AMD delayed the introduction of this trifle, and due to the lack of a thermal sensor, its processors burned out, which hardly contributed to its image. For that time, the protection was based on the principle of measuring the voltage on the processor’s thermal diode, and when it decreased below the threshold, the PS_ON signal for controlling the power supply was turned off (in Pentium2 it is somewhat different). One interesting point — on nForce2 motherboards there was no temperature measurement using the processor’s thermal diode, which is quite natural, but when you tried to connect it and change the type of monitoring input from «thermistor» to «thermal diode», strange results were obtained, the temperature was read from the diode by 25 degrees more . For calibration, I used a temperature measurement with the processor turned off, which guaranteed its temperature to be equal to the environment, but this offset of 25 degrees was preserved! (It can still be caused by interference and its detection on the diode itself, I got such a parasitic effect before installing the filter. ) And then it’s time to turn to the original source.
In order for the temperature to shift by 25 degrees, it is necessary to increase the current through the diode by 5 times. The nominal current through the diode is 0.1mA, the range is 5uA-0.3mA, AMD graphics were taken up to 0.3mA, which once again indicates the maximum permissible current. It follows from the output characteristics that the current through the diode is 0.5mA(!). From Theory to Schematic — In the Attansic ATTP1, the diode current is set by a resistor at the Vref (3.75V) output. The resistor value is 5.6KOm, the diode voltage is about 0.7V, it follows that the current will be (3.75-0.7)/5.6K=0.54mA. Why was it necessary to violate the specification so impudently, if it was enough to put a small resistor in series with the diode? There are no words.
The root of the problem was in the different thermal range of the AMD K7 and Intel Pentim2/3 processors, for which, apparently, ATTP1 was developed. If the latter had a low response temperature (presumably 70 degrees), then AMD had the sensor in a different, hotter place, which required a shift in the response threshold. The sensors themselves (diodes) in Pentium2/3 and K7 have similar characteristics. I really did not want to run MBM5 and correct the offset in the registers, I tried to reduce the current to the nominal value — the diode began to show quite correct readings, but also turned off 25 degrees earlier. If before revision it was 90-95 degrees, then after the return of the current to the nominal value (0.1mA), the shutdown occurred at approximately 65 degrees. For K7 with overclocking, getting this temperature is quite simple, so I had to add this unfortunate additional resistor.
Somewhat earlier, even before the Pentium2, PCs implemented a technology for controlling the frequency (more precisely, performance) of the processor through throttling. In Russian, this concept is closest to the definition of PWM. The consumption of the processor depends on its frequency, so if you periodically turn off its frequency (for a short time), the processor will heat up less. For example, if you turn off the frequency for 1 second every 2 seconds, then the processor will work only 50% of the time (out of 2 seconds one = idle), heat up only 50% of the usual value, but also do only half of the work. Of course, it hardly makes sense to switch PWM with a period of 2 seconds, there will be jerks in all programs, the equipment may work unstable, so they try to make this period small. PWM is not used as a processor performance regulator; its place is in thermal protection systems. When the processor overheats, then you have to choose — either turn off the computer, or try to somehow keep it working. Throttling was enabled and configured in the BIOS, was purely a software element, had low reliability and… and was quietly forgotten. In modern computers, this line is no longer there. The Pentium2 processor, with its built-in temperature sensor, made further progress in protecting the computer from destruction.
The protection technology is trivially complex — watch the voltage on the thermal diode and, if it drops below the threshold, turn off the power supply. The solution is completely hardware, not dependent on software problems. With a high science intensity, this action also gave positive results — the processors stopped burning out from the fan stop. But due to the fact that the temperature sensor in K7 is not at all at the hottest point of the processor, the response time of the protection turned out to be large and does not work if the cooling deteriorates sharply (for example, the cooler has fallen). For example, THG had difficulty with this.
A small quote from the review: «The Palomino thermal diode cannot respond quickly enough to temperature changes. If the temperature changes faster than one degree per second, the sensor will not be able to inform about it.» The reasons for this effect will be discussed below. Further, the development of overheating protection technology went in a fundamentally new and very original way — TM1 appeared in Pentium4. Its essence is that when the processor overheats, it exposes some hardware sign, and the frequency control PWM will turn on in the processor. Dejavu? Once I already met this method, for the life of me, I can’t remember when. (The only difference from trottling discussed earlier is that the throttling node has moved from the chipset to the processor). The effect was achieved, the processor was able to work even with the heatsink removed … though not too fast. This was shown very impressively in the video on removing the heatsink ‘on the go’ — Pentium4 continued to work, Pentium2 simply turned off the computer, K7 went into smoke. «AMD IS BURNING!» … on which Intel made quite good money, and, I think, this forced AMD to pay attention to flaws in the design of thermal protection of its processors.
Some time has passed, AMD has various technologies for dynamic frequency and voltage control of the processor PowerNow, C&Q and others. The essence was the same — to reduce the heat dissipation of the processor, it is necessary to reduce the voltage on it (consumption is proportional to the square of the voltage), and to reduce it, the frequency must also be reduced. A short time passes, and Intel has a similar technology, but, unlike AMD, designed not to reduce heat dissipation in idle time, but to protect against overheating. The idea is that when the temperature gets above the maximum, the processor reduces the frequency (not much, 2/3 of the nominal) and its voltage. In total, this leads to a 1.5-2-fold decrease in heat release. Not so much, but, unlike PWM, it does not generate jerks in performance. So the mechanism with the name TM2 turned out. TM1 and TM2 work from the same hardware sensor (sensors for a multiprocessor system), which cannot be adjusted. A separate sensor is made to measure the temperature, and it serves a more trivial purpose — controlling the speed of the fans. As far as I know, at the moment there are no official sensors in Intel/AMD processors that are allowed to be used to measure the temperature of the processor. Everything that is, goes as ‘only for testing’. Their accuracy is not guaranteed at all, register descriptions are removed from the documentation.
There is an interesting situation with Core2 — TM2 seems to be there, but it is not. For the protection mechanism to work, both frequency and voltage must decrease. The frequency (more precisely, the multiplier) is changed by the processor itself, there are no problems with this, but often nothing happens with the voltage. Nonsense? No, ‘developer’s conservatism’ — in the BIOS there is a setting for controlling the processor voltage — numbers, none, auto. Well, the numbers are clear to me, but what is the difference between none and auto? Auto — what will the BIOS itself think out and overclock? Hmm. In more detail, it looks like this — first, the voltage is set by VID, and then the song begins — the value of VID is overwritten by some value. When TM2 is running, the VID voltage changes, but the BIOS has rewritten the VID value to a constant, it cannot change. Those. the processor asked to reduce the voltage and ‘ignore’. TM2 and Enhanced SpeedStep are essentially synonyms, only the first works when overheated, the second when idle. This means that even in idle time, Core2 cannot cool down. AMD had something similar on K7, then a series of events followed, and for the next generation (K8), motherboard developers were forced to switch from rewriting a constant to adding an offset. As a result, a strange ‘+/- number’ processor voltage control appeared in the BIOS, but energy-saving technologies work.
What does Intel have? Hopefully this issue will be fixed in the near future. If the BIOS is set to ‘none’, then this will cancel the intervention of the ‘advanced developers’ of the motherboard, and the processor will control its own voltage. Because Core2 has a significant consumption in idle mode (20-50% of the maximum), and this number is weakly dependent on frequency (less than 1W with a decrease in frequency by 1.5 times), then voltage management will positively affect idle heat dissipation. By the way, Intel’s Enhanced SpeedStep is really fast, unlike the AMD version. The fact is that Intel has only one stage of the low mode, while AMD has several and very long ones.
And back to our sheep.
If you take the picture from the AMD document and put a heat sink, you get the following:
The numbers indicate specific points:
1 — the location of the thermal diode
2 — the point of maximum heating of the radiator
3 — the point of maximum heating of the crystal
4 (added) — the place on the radiator directly above the thermal diode.
According to AMD documentation, it turns out that 2.2 degrees fall between the maximum heating point (2) and the heatsink (3). I have no reason not to believe such a reputable company, but somehow I can’t believe either. Further, the temperature of the thermal diode (1) is 4 degrees lower than the maximum temperature for the core. Why below? It is necessary to dwell on this in more detail. The processor is not a monolithic resistor, when using different blocks it heats up, but the heating itself is uneven both in time and in the crystal. For example, according to Pentium4 Prescott, the thermogram looks like this:
We return to the previous picture. Point 4 is heated from the crystal directly below it (point 1) and due to the thermal conductivity of the radiator from point 2. The temperature sensor is located on the periphery of the crystal, there is never a significant heat release there. Firstly, look at the Pentium4 thermogram, and secondly, I have a suspicion that the sensor is not located in the active zone of the crystal. Along the periphery of the crystal there are contact pads, to which conductors are welded for packing into the case. For safety reasons and to reduce capacitance, a pocket is made under each pad — a p-n junction with respect to the substrate. On the processor case, the leads of the thermal diode are located side by side, it can be assumed that the contact pads are also nearby. Well, making a diode on two adjacent pockets is not difficult. Judging by the picture, the thermal diode is near the conclusions. Actually, there was no need to change the trace in the diode zone to move it by fractions of a millimeter. It is worth adding here that the early versions of K7 (T-Bird and earlier) did not have a diode, and the introduction of an extra component, albeit a small one, would require re-layout and re-routing of the crystal section from the pins to the sensor. If so, then this means that the thermal diode is at a very significant distance from the elements of the crystal that cause significant heating. Extremely exaggerating, we can say that the thermal diode heats up only due to heating from the radiator. (Very strong exaggeration, crystalline silicon has a significant thermal conductivity, about a third of copper). This means that the temperature at point 1 is not too hot at point 4, and the temperature at point 4 itself depends on the efficiency of the radiator. If the radiator has a thick base, small fins and is not blown, then the temperature at points 2 and 4 will not differ. And if you put a thin-bottom radiator with high fins and blow it through well? Then point 2 will be much, much hotter than point 4 (due to the rib between them, it will remove heat into the air and work like a transformer). This will result in the thermal diode will show a very low temperature . Earlier I gave a quote about the K7 sensor, now it’s time to explain its meaning — without a radiator (or with poor thermal contact), the thermal diode simply does not have time to heat up to the shutdown threshold due to the low rate of heat transfer through the crystal. And at this time, in the rest of the crystal, the temperature will have time to rise to a crazy value, and an irreversible process of destruction will begin. This behavior (slow reaction time) confirms my theory that there are no hot spots near the sensor.
According to K8…maybe that’s why it has such a thick lid? For example, in Core2 the sensor is located in the hot area (see the picture below), it does not need such tricks, and the lid is much thinner. The fact that the sensor is not in the hottest part of the processor means that the processor is using extrapolation. Accuracy of such extrapolation? … missing. I remember an anecdote about a drunk who was looking for keys by a lantern because it was lighter there.
At the very beginning, I talked about measuring temperature with a thermistor and gave the thesis that it can be more accurate than the sensors built into the processor. I think now you understand why it is potentially more accurate — it is pressed against the ‘belly’ of the processor and measures actual average processor temperature. In addition, loading with different programs means using a different set of processor blocks, which will cause uneven heating across the chip, which will affect the thermal diode (due to its local location), and not significantly for the thermistor due to the averaging effect. Somehow I compared the nature of the readings of a similar thermistor and a thermal diode in K7 (Barton) — the behavior is similar, the thermal diode has a clearer reaction to heating, but the graphs themselves coincide.
Unfortunately, I have no data on the location of sensors in modern AMD processors, but judging by the nature of the behavior, the technology has not undergone fundamental changes in the transition from K7 to K8. Intel is a little more certain:
There are two types of sensors in Core2 — a conventional thermal diode for the operation of the fan control circuit and two sensors (according to the number of cores) for thermal protection (DTS, Digital Thermal Sensors). The sensors themselves are located in the most heated parts of the crystal and reflect the real temperature, but for this crystal location . The temperature of the processor cover will be significantly lower for the same reasons as AMD — the final thermal conductivity of silicon and the thermal interface. The advent of Core2 processors has caused a new wave of interest in measuring the temperature of the processor. On the one hand, the meters are already built into the cores, which seems to guarantee their accuracy, on the other hand, some strange data was observed. Promised a cold processor with a high performance-per-watt ratio; the radiator is cold, and the sensors show crazy temperatures! Everyone is already accustomed to 50-60 degrees for AMD K7/K8 and Intel Pentium4, and this is such a frank ‘disgrace’. If you have read the text before, then you understand why the sensors on the chip can show such different data — it all depends on the location.
There is an interesting moment with the DTS itself — for the overheat protection system to work, the current temperature is not needed at all, it is enough to have a ‘countdown’ until the processor is turned off. Intel did just that — it is this very count that is measured. When 3 degrees remain before ‘0’, the thermal protection is activated. Why not at 0? — apparently for a margin for changing the aggressiveness of the processor cooling. When DTS reaches 0, this means that the resources to reduce processor consumption are exhausted, only stars (more precisely, sparks, smoke and soot) are farther. Why care about the accuracy of the protection triggering threshold if the processor will never run into it (as was the case with Prescott)? The processor is cold, more precisely — it was originally designed as a mobile one, therefore it cannot get very hot. There is no need to provide a clear threshold, and they do not adhere to it. As shown earlier, the temperature change is linear and hopefully the processor designers were able to correctly translate ‘-2mV/C’ into temperature. But the Vo offset … that’s the problem with it. But there was no particular need, and they did not complicate their lives, this parameter is compensated quite relatively. When Core2 Quad appeared, it immediately became noticeable that the temperature may differ for different crystals. Moreover, the temperature change of the nuclei from heating was the same, only the absolute numbers differed. I have no statistics, but a difference of 3-5 degrees in C2Q is a common situation. Another problem, and already far-fetched, is related to the operation of temperature reading programs. Because DTS shows a countdown, and not an absolute temperature, then to convert to a generally accepted format, you need to know the temperature limit (Tjunction). Again, this is the temperature when the overheating protection mechanism was triggered. For mobile versions of the Core2 processor, information has leaked that there is a bit in some MSR that declares a Tjunction of 85 or 100 degrees. Op! Let’s apply this to the «non-mobile» versions of the processors — here are the Core2 core temperature display utilities. In order to shorten the time a little, let me quote
Intel Software Network Support Specialist dated March 2003:
Use of bit 30 in MSR 0EEh is *not* valid for desktop, workstation or server processors based on the Intel(R) Core(TM) microarchitecture. However, for mobile processors, this assumption *is* valid.
So the switch 85/100 is incorrect, but which of the two numbers is correct? …or are both wrong? When I wrote the Core2 temperature measurement module, I was immediately surprised by the strange data on processor temperatures. On the one hand, e4300 and e6600 indicated Tjunction at 85 and 100, which is a priori stupidity (processors of the same family cannot differ so much). On the other hand, at Tjunction=85 the temperatures seemed to be normal and matched the third processor sensor that calibrated the BIOS. But Tjunction=100 produced something unintelligible. Once there is a problem, it must be solved, and a practical test is best. In order to reduce heat loss at the transition through the thickness of the silicon sensor-cover and cover-radiator, I reduced all voltages and frequencies to a minimum and turned off the fans and turned off the computer for a long time — so that the processor cooled down to room temperature. Then I turned it on and watched the temperature increase on the thermometer and DTS. The processor emits little heat, I had to help it a little with additional heating of the water (I have a CBO). In the range of 25-60 degrees, there were no significant differences in the change in water and DTS temperatures, the shift between them remained almost unchanged. If we take Tjunction = 100 degrees, then at DTS it was 4 degrees higher than the water temperature. From here, the option with Tjunction = 85 degrees immediately disappeared — I do not have a refrigerator in the system 🙂 .
Although the processor was in a low-power mode, it consumed ‘something’ (about 7 watts), so the DTS readings should be slightly higher than the temperature of the lid. Another argument in favor of the option with Tjunction = 100 degrees — mobile versions of Core2 are now being released and Tjunction = 100 for them.
correct core temperatures, and do not emulate who knows what. By the way, on emulation: I recalculated the temperature of the cores according to DTS to the estimated temperature of the lid. The numbers are not so defiant, but are absolutely wrong. 🙂
On the other hand, if you wish, you can calibrate the sensor-cover temperature difference compensation, then you can relatively correctly measure the efficiency of cooling systems through DTS (it won’t work on AMD, the heating and the sensor are in different places). The area is not plowed, for example — silicon has a non-linear thermal resistance, especially in the upper thermal range (70-90 degrees).
By the way, AMD is also not going well with measuring core temperatures. At first, a temperature meter was declared, then a temperature offset appeared to correct it for a specific instance (see above about Vo). The idea is correct, only the offset data is in the same place where completely different information was previously stored. As a result, when outputting the corrected temperature, some nonsense was obtained (K8 Venice). At home, I refused to make compensation and give the original temperature. Do not be surprised if different programs show slightly different core temperatures on K8, this may be the source of the problem.
If we slightly paraphrase a well-known saying, then only one thing can be said from the software temperature reading — there are lies, blatant lies and software monitoring.
Trust nothing, measure with your finger or use Intel’s method:
«To measure the operating environment temperature of an Intel® processor, place the temperature probe 0. 3″ above the fan inlet. If the fan is not reversed, air should be expelled. If the fan is blowing air out, measure the air approximately 0.3″ from the heatsink, closest to the inlet area. The outside temperature of the active cooling system is measured near the processor in the following ways. An active cooling system involves the use of a fan.»
I wanted to show the dynamics of changes in core temperature difference in Core2 Duo, but something else came up.
The upper half shows [green>CPU<] and [yellow>CPU frequency<], while the lower half shows temperatures by [red>DTS0<] and [purple>DTS1<]. The processor was transcoding 7 films using a different set of filters (avisynth) offline, no one touched the computer. It is clearly seen that the temperature of the cores floats relative to each other even with approximately the same processor load. This is logical, filters can use different sets of commands.
I was about to put an end to it, but I accidentally noticed that in the middle of the graph there is obvious stupidity (highlighted by a frame). Within 15 minutes, the frequency of the processor was underestimated. It can be blamed on an error in the frequency measurement program, but the processor temperature also decreased by 5 degrees in this time interval.
At the same time, the temperature itself was much lower than Tjunction, and the processor itself was loaded by 90-100%.
Well, not everything is in order in the Kingdom of Denmark.
The monitoring file is available here.
Subscribe to our channel in Yandex.Zen or telegram channel @overclockers_news — these are convenient ways to follow new materials on the site. With pictures, extended descriptions and no ads.
what is it and what temperature should it have
They say that the main thing in a gift is not the content, but the packaging. And in the processor? Logically, the most important thing in it is the cores, but why then is the CPU Package indicator so important for the system?
We recently figured out what the mysterious string «Diode PCH» means in hardware monitoring programs, and today we’ll talk about its neighbor. So, what is the «CPU Package» and what temperature should it have.
- What, where, why…
- And how many degrees?
- The temperature of the CPU Package is much higher or lower than the chip temperature. What’s broken?
What, where, why…
The central processing unit is the main and hottest unit of the system, lined with sensors on all sides. They are located both inside and outside.
CPU Package is one such sensor that is integrated into the processor itself. Its name, as it were, hints that it is related to a package — a socket (in AMD specifications, a socket is referred to as a package) or a heat-distributing cap of the CPU — that «piece of iron» on which processors for desktop systems are marked.
In fact, the packaging of the processor is not called the cover and connector, but the case — a part covered with green varnish, on one side of which there is a silicon crystal — the main computing node, and on the other — an interface for connecting to the computer motherboard (legs, balls or pads). The CPU Package is a thermal diode that detects the temperature of the CPU package in real time.
And how many degrees?
CPU Package in hardware monitoring programs is one of the main indicators of processor temperature. Its value, as a rule, is close to Core 1…n — core temperature. However, it is not surprising, because these elements are located nearby and are very closely related.
The screenshot below shows the temperature values of a dual-core mobile Intel Core i3 in the utility HWiNFO32/64 :
0186 AIDA64 :
It can be seen that the temperature of the CPU Package is equal to the temperature of the cores. These are normal indicators of a working device.
CPU Package is sometimes considered the main value to look at when determining the overall temperature of the processor. On the one hand, this is correct, but on the other, it is not quite right. This is correct only when the CPU Package and Core 1 . .. n have approximately the same indicators, as in the screenshots above (the difference of 1-5 degrees is not significant). And if the difference reaches twenty degrees or more, then the maximum temperature of Core 1 … n cores should be taken into account.
Why? But because the cores are controlled by Intel’s proprietary technology DTS (Digital Thermal Sensors), which protects the processor from thermal damage. When the chip temperature reaches the upper threshold, Intel DTS activates clock throttling or thermal throttling — reducing the processor frequency to a level at which its temperature indicators begin to decrease. She is also “guilty” of turning off the computer if the CPU overheating becomes critical.
A similar technology is implemented in AMD processors.
The normal temperature of the CPU Package is equivalent to the normal temperature of the cores of the crystal.
And the latter depends on the type and generation of the CPU. To summarize, the processors of desktop systems with good cooling and moderate load heat up to 45-55 ° C, mobile — up to 55-65 ° C with short-term rises to 70-85 ° C. You can read more about this in this article.
Long-term, and even more so constant heating of the processor by 10-20 ° C above the average norm affects the operation of the system unfavorably. This does not damage the CPU, but it can:
- To freezes, spontaneous computer shutdowns and, as a result, to the loss of unsaved files and operating system errors.
- To failure of hard drives. For them, not only overheating is dangerous (such a state of the processor is often a sign of overheating of the entire system), but also a sudden power outage.
- To the failure of the VRM elements — the processor power supply system, since they are also forced to work in abnormal conditions.
The temperature of the CPU Package is much higher or lower than the chip temperature. What’s broken?
Why do the CPU Package and Core show different temperatures if the package and the die are about the same temperature? When you see a picture like the screenshot below (made in HWiNFO), the first thing that comes to mind is a breakdown, because the CPU case cannot be 2 times hotter than the cores.
Of course not. Such implausible differences are usually due to incorrect interpretation of data by monitoring utilities, since information from sensors is processed by different controllers using different algorithms. This is not uncommon on AMD platforms.
Another reason for such curiosities is a malfunction of temperature sensors or elements of the control system. It is less common.
Differentiating utility errors from hardware failure is quite simple: incorrect reading of data by programs does not affect the operation of the computer in any way, and the failure of sensors and controllers always has external manifestations. For example, the system will signal overheating, although in fact it is not. Or suddenly turn off, which is fraught with even more serious consequences.
Finally, the rarest case is a processor defect, which manifests itself in the fact that the temperature of the CPU Package is much lower than the Core, and not nominally, but in fact. Such devices do not live long, as the defect tends to progress. But this is really very rare. It is much more likely that the program is trying to deceive you again.
CPU Package — although not the most important indicator, it can tell a lot about your system.
Good luck with your research!
CPU thermometer: an overview of temperature monitoring utilities
We all love summer for the long-awaited opportunity to bask in the gentle sun. However, this is a really tough time for our PCs due to an insidious enemy called overheating.
As you know, all computer components, and especially processors, emit a certain amount of heat during operation. For the normal functioning of the system, it must be removed, ensuring optimal temperature conditions. This task becomes much more difficult with the arrival of the summer heat, when standard COs can no longer cope, especially if you have been overclocking your PC.
The result of overheating can be either a complete failure of the computer, or permanent system instability, expressed in the appearance of strange errors, freezes and sudden reboots. In order to prevent such unpleasant consequences, it is necessary to keep the temperature of PC components under control using special programs, an overview of which we bring to your attention.
- 1 Core Temp | www.alcpu.com/CoreTemp
- 2 SpeedFan | www.almico.com/speedfan.php
- 3 Real Temp | www.techpowerup.com/realtemp
- 4 Hmonitor | www.hmonitor.net
- 5 CPU Thermometer | www.cputhermometer.com
- 6 PC Wizard | www.cpuid.com/softwares/pc-wizard.html
- 7 Open Hardware Monitor | openhardwaremonitor.org
- 8 Temp Taskbar for Windows 7 | anonymous-thing.deviantart.com
Core Temp is a compact, lightweight real-time CPU temperature monitoring tool. The utility works with both Intel and AMD processors, and a very wide range of models is supported, including the latest designs. CoreTemp can display in the system tray the temperature, frequency, or load for each core. In addition, a special gadget is available for download that displays these parameters on the Desktop, as well as a plug-in for Windows Media Center. The application has an overheat warning feature in the form of a tooltip.
SpeedFan is one of the oldest and most popular programs of this type. Its capabilities go far beyond simple temperature monitoring: it is able to read voltages, frequency, fan speed, SMART data, hard drive temperature, as well as display the current values of selected sensors in the tray. The main feature of the application is the ability to adjust the fan speed depending on the temperature of the processor. In addition, the utility maintains statistics and saves information to a log file, draws graphs of changes in temperature, voltage, and cooler rotation speed.
Real Temp is a temperature monitoring program exclusively for Intel processors (Core, Dual Core, Quad Core and Core i7). Works under all versions of Windows and does not require installation. Data is read directly from the sensor located in the CPU core, so the readings of the utility, according to the developers, are closest to real ones. Real Temp has the function of warning or emergency shutdown of the computer when the limit values are reached. With the application, you can also test the temperature sensor and CPU speed.
Hmonitor is able to monitor CPU and HDD core temperatures, voltages and fan speeds. The program has a wide range of options for setting the display of controlled parameters. The choice of various actions performed by the utility in case of processor overheating will allow you to automatically respond to extreme situations. This could be beeping, automatically shutting down the system, or running an entire script (for example, sending an email to an administrator). Hmonitor is paid (shareware) and can be used in all versions of Windows.
This program is probably the easiest to review. All she can do is display temperature and CPU usage in the system tray. The utility is a heavily truncated version of the Open Hardware Monitor discussed above, in which only one function was left. In the settings, it is allowed to specify the color of the indication and autostart along with the system startup. However, such limited functionality had little effect on the appetites of the application — this thermometer is much more voracious than its much more advanced competitors. Despite the free status, CPU Thermometer is unlikely to be the best choice.
PC-Wizard is a whole software package that can control and diagnose almost any computer components. This application can be used to analyze system configuration and test hardware, including CPU performance, RAM, hard drive speed, and more. When minimized, PC-Wizard can display the main system parameters, including the CPU temperature, in the form of a small display located on top of all windows. Thus, you can always see them and respond quickly.
The program is a product of the open source developer community. It monitors temperature sensors, fan speeds, voltages, load and clock speed of the computer’s processor. Open Hardware Monitor supports most modern motherboards for Intel and AMD platforms. In addition, the utility can read SMART readings from hard drives and video adapter sensors. Any parameter can be displayed as a special gadget on the desktop or as a tray icon. The application is used not only in all versions of Windows, but also in Linux.
The program is designed purely to show the temperature of the processor, but it does it in a very unusual way. After starting the Temp Taskbar, the entire Windows Taskbar becomes an indicator, on which a green CPU temperature scale appears, and the application buttons located on the panel do not interfere with its display at all. Upon reaching a high level, the scale turns yellow, and at critical values, it turns red. The corresponding parameters and colors are set in the utility settings. It is controlled by an icon in the system tray.
Video card temperature — how to find out: 3 monitoring programs
It seems that something is wrong with the computer: games, video — everything is dull, although the power of the components of the system unit is more than enough? Perhaps the whole thing is overheating of the video card. How to find out if the GPU is heating up and how to avoid overheating, the article tells.
Why know the temperature of a video card?
There is a recommended temperature range in which the GPU operates stably. Most modern components are equipped with protection, and therefore, if the temperature reaches a critical value, the system turns off. But any automation can fail, and therefore it is very important to check for overheating of components, including the graphics processor.
Unreasonably high temperatures, regardless of the load, indicate a malfunction of the video card. This can cause not only incorrect operation, but also failure. «Heat» inside the system unit can negatively affect the operation of neighboring components. That’s why you need to know the optimal temperature of the video card. What is the optimal one? This is the next section.
Interesting: How to assemble a gaming computer, what components to buy: 3 options to choose from
What temperature is considered normal for a video card
Different video adapters have different normal temperatures. But there is still a certain threshold, above which you should pay attention to the state of the GPU. More on this later. In the meantime, it is worth figuring out what affects the temperature of the video card.
The main factors that affect GPU t°:
- The processor on which the device is based. Each GPU generates a certain amount of heat (the indicator is displayed in the characteristics as a DPI value).
- The type of cooling system is passive, which uses only a radiator, and active, which provides for the presence of one or more coolers.
- Cooling of the system unit itself. The boxed cooler that comes with the processor is not always able to provide the necessary cold. In addition, the number of “turntables” (for blowing and blowing) and their diameter are important.
As for optimal temperatures, the range of 30-60 degrees in idle mode is considered acceptable for most models. The critical mark may also differ. Some models have a limit of 80 degrees, while others have a limit of 90°. There is a separate section for this.
How to find out the temperature of the video card
The normal t° of the video adapter is usually indicated in the technical specifications or instructions. You can check if it exceeds the norm yourself. Alas, there are no tools integrated into the operating system that will help determine the temperature of individual computer components. To find out if everything is in order, you will need to go into the BIOS.
Important : if the user is not confident in his abilities, it is better to entrust the verification to a professional.
It is worth noting that checking in this way will not help to know the temperature of the components during load. In addition, the state of the GPU is not always displayed in the BIOS, and in this case, third-party software is indispensable. Good monitoring programs are discussed in the relevant section below.
Causes of high temperatures
Why can a video card get very hot in general? The most common causes of overheating are not so many:
- Failure of the video card cooling system or PC coolers.
- Fault in the motherboard: if it heats up, then the temperature inside the case also rises.
- Housing dusty.
- Malfunction or low power of the power supply.
- GPU thermal paste dried out.
Permissible and critical values for video cards of different brands
Intel HD is an integrated video chip, so its temperature depends on the CPU temperature. Factors such as model, power, operating mode (whether the CPU is working in normal mode or overclocking), etc. influence the optimal and maximum value, etc. In one case, rising to 75 degrees can cause an automatic reboot of the system, and in another — and at 100 degrees the processor will continue to work.
This does not mean that long periods of operation at peak performance will not do harm. At maximum loads, the device works at its limit, which means that it is necessary to correct the situation.
80-90 degrees is not a cause for concern if the heating occurs at the maximum load of the video card. Indicators above 90-95 degrees, especially with prolonged use, can already harm the device, despite the fact that, according to the manufacturer, the peak mark can reach 100-105 degrees.
According to the manufacturer, devices can withstand up to 105 degrees, but it is not recommended to test the video card for endurance. 80-90° — acceptable limit: depending on the model of the video adapter, when the mark exceeds these indicators, the device becomes unstable.
Note : 90-95 degrees for AMD/ATI and NVIDIA products may be acceptable during peak load or overclocking.
Selection: 3 best manufacturers of NVidia GeForce GTX 1060 graphics cards + 5 cool models
What to do if the video card is heating up?
Disable GPU overclocking
In an attempt to get the most out of the video card, even if you use proprietary software from the kit, you can overdo it. At the first sign of overheating — unstable operation of the video adapter — it is recommended to return the default settings.
Clean fan and heatsink
Regular cleaning of the cooling system, and indeed the case in general, allows you to maintain comfortable conditions for the operation of the GPU. Dust that gets into the radiator or fans accumulates and eventually clogs the system, preventing it from functioning normally.
Thermal paste replacement
Compound that improves heat transfer between the heating element (CPU, GPU) and the heatsink tends to dry out. Its replacement once a year, and preferably every 6 months, helps to stabilize the cooling process of the device.
If the fan on the video card or in the system unit itself does not work, this becomes a serious problem: components overheat and quickly break down. Other than contacting a specialist to repair or replace coolers, nothing can be done.
Install a more efficient cooling system
Perhaps the cooling was chosen incorrectly, its efficiency has decreased due to wear, or power has become insufficient after upgrading the computer. All that can be done is to replace the system. This can be water or more powerful air cooling, installing coolers with a larger diameter, or adding additional fans if the case allows.
Curious: Comparison of 2 budget gaming graphics cards — GTX 1050 vs GTX 1050 TI
Programs for determining the temperature conditions of the graphics adapter
As mentioned above, monitoring the temperature of the video card and other PC components is easier and more convenient using special applications. Below are three of the most popular utilities.
The software is convenient and easy to use. With it, you can find out all the necessary information about the video card. In addition to the temperature of the device, the user can find out the model of the video adapter, the amount of memory, the load level, frequencies and other parameters.
The program is well known and popular. With this utility, you can get comprehensive information about all system components. To find out the temperature of the GPU, the user only needs to open the “Sensors” tab.
Another handy program for diagnosing and testing computer assembly components, with which you can monitor the status of devices. The lack of a Russian-language interface is perhaps the only drawback of the software, but the software itself is very user-friendly: you can intuitively figure out what’s what.
Read also: Which video card is better — NVidia GeForce GTX 1050 or AMD Radeon RX 560
This indicator depends on the specific model of the video adapter, and on the manufacturer, and on other factors. However, both the optimal and the critical value are indicated in the specifications of the models. You can find out if the card is overheating in different ways: through BIOS and through special programs. The second option is better: there will be no chance of pressing something wrong and negatively affecting the operation of the system. You can also prevent overheating of the component. It is enough to monitor the condition of the PC: clean the case from dust in time and change the thermal paste.
CPU temperature on a computer in Windows: what should it be and how to find out?
It’s no secret that a computer’s processor is its key component. After all, it is the brain of the entire system, performing many operations, even when it seems to be sleeping. And the main enemy of any computer hardware is temperature. The more and more intensively the device works, the more heat it generates. Because of it, the processor can even significantly reduce its performance. But what should be the temperature of the processor, how to monitor and limit it?
♥ RELATED: How to clear the cache on a Windows 10 PC (and DNS cache) without installing additional software.
Why does the CPU get hot?
As we have already said, the number of tasks assigned to the processor and their complexity directly affect the heating of this unit. And you can see the load level of the main computing module of the computer in «Task Manager» . To call it, in Windows, press the key combination Ctrl + Shift + Esc. In the very first line of tab «Processes» («Performance» for older versions of Windows) will show the total CPU usage, and the following lines will indicate which application consumes how many resources.
The graph (Performance tab ) shows the history of CPU usage over the last minute. If the increase in processor temperature is associated with a load on it, then this is normal.
The same load type can load the processor differently. The top multi-core representatives of the Core and Ryzen families can be loaded with a few percent of the tasks that take 30-50% of the resources from the Celeron. Yes, and background tasks that the processor is engaged in idle state can seriously load the system.
In addition, the causes of processor overheating can be: installation of malware, incorrect operation of temperature sensors, insufficient fan speed due to dust accumulation, and dried thermal paste.
♥ BY TOPIC: How to clear (delete) Telegram cache on iPhone, iPad, Android, Windows or Mac computer.
How to detect overheating of the processor?
An increase in fan speed and resulting noise, as well as an increase in power consumption by the device, may indicate an overheating of the processor. And accurate information about the temperature of the processor can be provided by the BIOS (UEFI) or system utilities.
♥ RELATED: How to Fix Windows 10 Errors and Improve Performance: Best Programs (Free). Core Temp It displays basic information about the processor, including its temperature. Core Temp works with all current versions of Windows, including Windows 7/10.
The program allows you to find out the temperature of each processor core and displays this information in the Windows taskbar. If Core Temp is in the system startup, then the necessary information will always be present in the taskbar. The program can also be used as a data provider for the All CPU Meter desktop widget. But Core Temp also has its own widget designed for Windows 7 — Core Temp Gadget. Another interesting addition to this program is Core Temp Grapher, which plots load and CPU temperature. Both plugins can be found on the add-ons page of the official Core Temp website.
CPUID HWMonitor is a free application that is considered one of the most popular for viewing the status of the hardware components of a computer or laptop. HWMonitor also shows the temperature of both the processor as a whole and all its cores individually. Moreover, depending on the model of the motherboard, it will be possible to find out the temperature of the socket. The Min and Max columns show the average values, while the Value column shows the actual temperature.
HWMonitor can also show useful information such as:
- The temperature of the video card, hard drives and motherboard.
- Rotation speed of various fans.
- Information about the voltage on the components and the load on the processor cores.
Download HWMonitor (Free)
This simple program is great for beginners. The Russian-language application provides information about the main devices of the computer and their characteristics. In addition to a lot of interesting and useful information in Speccy, you can also see important readings from the temperature sensors of your computer or laptop. The CPU section will show the current CPU temperature. In other sections, you can find out the temperature of the video card, motherboard, SSD and HDD drives, if they have the appropriate sensors.
Download Speccy (Free for non-commercial use)
This program is commonly used to monitor the fan speed of a laptop or computer cooling system. But at the same time, the utility can also provide information about the temperature of the main components of the system, from the processor and its cores to the video card and hard drives.
The program is constantly updated, reading data from all current motherboards. It works correctly in both Windows 7 and Windows 8/8.1/10. But on older systems, SpeedFan can cause problems when trying to adjust the fan speed. There are also several additional features in the program. For example, she can plot temperature changes. This can be useful to see how hot the processor gets when performing resource-intensive tasks.
Download SpeedFan (Free)
♥ RELATED: 10 Unnecessary Standard Programs in Windows 10 You Can Remove.
Normal processor operating temperature — how many degrees?
This question is often asked by novice users, and it is relevant for both laptops and computers, both for Intel-based and AMD-based solutions.
There is no universal answer to this answer. Even two processors of the same index under the same conditions can have different operating temperatures. The difference will be small, but still noticeable. In the description of processors, the phrase «maximum operating temperature» is always present. This does not mean at all that work at indicators below critical will be normal. A temperature of 100 degrees can be considered fatal for most processors. But in fact, the system will begin to protect its “brain” even at lower rates, including throttling.
Under this name, the processor protection system from overheating is hidden. As the temperature rises, cycles will be skipped. With a decrease in load, the heating of the module will also decrease — it will return to the previous mode. In general, the recommended operating temperature is 70-80% of the maximum. High temperatures will reduce the life of the processor. And to work in the optimal temperature range, you need an efficient cooling system. It is believed that at low load, the processor should not heat up over 35-50 degrees.
Actual solutions from Intel today (this includes most models released over the past 10 years) operate within the following temperature limits:
- 28 — 41 degrees in idle mode. It implies a running Windows desktop, but no system maintenance operations are performed in the background.
- 40 — 62 degrees (50 — 70 degrees for models overclocked to high frequencies) in load mode. This includes demanding games, video rendering, archiving tasks, working with virtual machines, etc.
- 67 — 72 — maximum operating temperature, according to Intel recommendations.
For AMD processors, temperature standards are about the same, except that for some representatives of the FX series, the maximum recommended temperature is only 61 degrees Celsius. Do not be afraid if in load mode your performance is slightly higher than recommended. This may be if the computer or laptop has been working for a long time.
♥ BY TOPIC: How to automatically switch the language (layout) on the keyboard in macOS or Windows.
What to do if the processor overheats?
If you think the processor temperature is too high, then you can try to reduce it in several ways:
1. The free SpeedFan utility will help speed up the rotation of the fan blades on the processor. This will remove excess heat and lower the temperature.
2. Open the system unit cover. But this solution is temporary, and the components will be unprotected from dust.
3. With the lid open, you can try to remove hot air with a vacuum cleaner or a fan.
4. Replace the thermal paste on the processor.
5. If you have a passive cooling system, you should replace it with an active one. Water cooling will show the greatest efficiency.
6. Install additional case fans. The temperature inside the case also affects the temperature of the processor. The faster cold air gets inside the system unit, the better.
7. Adjust the speed of the CPU cooler depending on the load. But setting 100% rpm on a permanent basis is not worth it. If the cooler does not cope, you can change it to a more powerful one.
8. If you constantly work with programs that heavily load the processor, try replacing it with a more productive one.
- Hidden folders and files in Windows 10: how to hide (show) photos, videos, documents, drives.
- How to speed up your Windows 10, 8, 7 computer: 20 tips.
- Windows «con» folder: why can’t a directory be created with that name?
Please rate the article
Average rating / 5. Number of ratings:
No ratings yet. Rate first.
Normal CPU and GPU temperatures in games
It’s always worth remembering that a gaming PC needs regular maintenance. Many of us have learned this truth from our mistakes, especially those who started playing at an early age and knew almost nothing about computer hardware. And when we encountered that eerie blue screen of death, it was almost certainly caused by overheating.
It’s common these days to take care of your expensive gaming machine, and the best way to know if your PC is getting hotter than usual is to monitor component temperatures with built-in sensors.
That’s what we’re going to talk about in this article: how to monitor the temperature of the processor and video card, what is the optimal temperature for gaming, and what could be the reason for the increase in temperature.
Optimal temperature for gaming
Today, acceptable temperatures are a little lower than they once were, mainly due to the more sophisticated manufacturing technologies of today’s processors. These levels differ depending on the manufacturer and even the specific model, so it is simply impossible to specify one exact value for the ideal temperature for games.
Next, we will look at the allowable temperature ranges for the various components.
Currently, most gamers are faced with a choice between two series of processors: the Intel Core and AMD Ryzen lines.
According to manufacturers, the maximum safe operating temperature is 95 and 100 degrees Celsius for Ryzen and Core processors, respectively, but you are unlikely to encounter such values if the cooling system is working and the processor has not been overclocked beyond the cooler’s capabilities.
Even under heavy load, the temperature of Ryzen or Core processors should not exceed 85 degrees with a stock cooler without overclocking. If it is higher, this is a sign of some problems, but we will talk about them later.
There are also two main manufacturers of video chips. This is NVIDIA and again AMD. However, these companies only develop and manufacture video processors themselves (not counting the reference video cards from NVIDIA), and most video cards are made by other companies, such as Asus, Gigabyte, EVGA, MSI or Sapphire. They implement their own cooling solutions.
But this does not change the maximum safe temperature setting for NVIDIA GeForce and AMD Radeon cards. In both cases, the upper limit is around 95 degrees, although Radeon cards can withstand higher temperatures due to AMD’s more robust video chip architecture.
Like processors, most video cards should not get hotter than 85 degrees even under heavy load. Of course, as mentioned above, the quality of the coolers used by the video card manufacturer will affect the average temperature, and cheaper models with the same video chip usually heat up more than more expensive ones.
There are two main types of air cooling for video cards:
- Fan — the most common type of cooling for video cards, in which one, two or three fans drive air through an open heatsink. Efficiency is greatly enhanced by good air circulation in the case and the presence of case fans to help expel hot air from the case.
- Turbine — is found on most reference models and is not as popular. In this case, the video card is completely closed, and the only fan sucks in cold air and blows hot air out through the holes in the back of the card. These cards typically run hotter than fan-cooled models and are only really preferred when space is limited inside the case and/or air circulation is poor.
But there is one more nuance here: most modern cards (if not all) use adaptive ventilation technology. What does it mean?
In fact, when using this technology, the fans do not turn on until the temperature reaches a certain threshold, most often around 40-50 degrees. This was done to reduce power consumption and fan noise when the card is not under load. Therefore, when the fans are turned off, the card may seem hotter than it should be.
As in the case of processors, a non-overclocked card with working fans should practically not heat up above 80 degrees.
Video chip Temperature Nvidia GTX 950 95°C Nvidia GTX 960 98°C Nvidia GTX 970 98°C Nvidia GTX 980 98°C Nvidia GTX 980 Ti 92°C Nvidia GT 1030 97°C Nvidia GTX 1050 97°C Nvidia GTX 1050 Ti 97°C Nvidia GTX 1060 97°C Nvidia GTX 1070 94°C Nvidia GTX 1070 Ti 94°C Nvidia GTX 1080 94°C Nvidia GTX 1080 Ti 91°C Nvidia RTX 2070 89°C Nvidia RTX 2080 88°C Nvidia RTX 2080 Ti 89°C Nvidia Titan X 94°C Nvidia Titan V 91°C Nvidia RTX 3050 93°C Nvidia RTX 3060 93°C Nvidia RTX 3060 Ti 93°C Nvidia RTX 3070 93°C Nvidia RTX 3070 Ti 93°C Nvidia RTX 3080 Ti 93°C Nvidia RTX 3080 Ti 93°C Nvidia RTX 3090 93°C AMD RX 460 64°C AMD RX 470 75°C AMD RX 480 80°C AMD RX 560 62°C AMD RX 570 74°C AMD RX 580 72°C AMD RX 590 78°C AMD Vega 56 75°C AMD Vega 64 85°C AMD Radeon RX 6400 80°C AMD Radeon RX 6500 XT 85°C AMD Radeon RX 6600 87°C AMD Radeon RX 6600 XT 88°C AMD Radeon RX 6700 XT 88°C AMD Radeon RX 6800 88°C AMD Radeon RX 6800 XT 88°C AMD Radeon RX 6900 XT 88°C
For Nvidia graphics cards, the temperatures listed as the maximum safe on the respective official pages of the Nvidia website.
The values given for AMD cards are average temperature levels for heavy workloads and can be used as a fairly reliable approximation of how hot your graphics card is when working hard.
How to monitor the temperature of components
Thanks to a huge number of sensors built into processors, video cards and motherboards, you can see exactly how hot each element is. But what programs are best for this?
The easiest way to check the temperature and other current parameters is through the BIOS. Just restart your PC and enter the BIOS by pressing Delete during boot.
However, the obvious disadvantage of using the BIOS is the need to restart the PC. In addition, for obvious reasons, it will not be possible to monitor the temperature directly under load. However, if regular monitoring is not required, this method is the most convenient, since it does not require any additional programs.
Processor and graphics card utilities
Intel, Nvidia, and AMD have handy utilities for working with processors and graphics cards.
For processors, these are Intel Extreme Tuning Utility and Ryzen Master Utility . Both programs allow you to find out a wide variety of parameters, easily overclock the processor and, most importantly, see the current temperature of the processor.
As for graphics cards, there are overclocking utilities for them that can also be used for temperature monitoring: MSI Afterburner, Asus GPU Tweak, Gigabyte Aorus Graphics Engine, etc.
Third party programs
There are many different programs for temperature monitoring, but we recommend one of these two: OpenHardwareMonitor and AIDA64.
- OpenHardwareMonitor — is a completely free utility that allows you to monitor many important parameters: temperature, voltage, fan speed and many others.
- AIDA64 – is a very popular and incredibly powerful utility with outstanding functionality including temperature tracking. But it’s not free, so you’ll have to buy it or use the trial version.
Does temperature affect performance?
You are probably wondering if your computer runs faster at lower temperatures and vice versa, slower at high temperatures.
If the temperature is kept within an acceptable range, you will not experience performance drops. But at high temperatures, the so-called throttling begins: the processor (CPU or GPU) resets the frequencies to lower the temperature, and this, of course, affects its performance. So, if you want to achieve maximum performance of iron, take care of cooling.
What to do in case of overheating?
Overheating of the processor and video card can be caused by several reasons:
- Heat sink contamination
- Poor air circulation in housing
- High ambient temperature
- Malfunction of the cooler, power supply or the processor / video card itself
Here’s what you can do:
Clean the heatsink(s)
If you’ve had your computer for over a year and never cleaned it, a lot of dust can build up inside. You can easily get rid of it yourself, with a brush and a vacuum cleaner.
Check Air Circulation
One reason for excessive heat build-up in components can be poor air circulation—when the CPU and GPU coolers don’t get enough cool air to move through the heatsink. If this is indeed the reason, the motherboard sensors will also show a higher temperature. Also, check your CPU cooler and make sure it’s clean.
The best way to improve air circulation is to set case fans for intake and exhaust. But before that, you should familiarize yourself with the principles of air circulation in order to create optimal airflow inside the case.
If you can’t or don’t want to buy additional fans, you can always keep the case open. This will significantly reduce the heat.
Check the ambient temperature
This problem is mainly for those who live in hot climates, although even residents of more temperate regions may experience it during warm summers.
Unlike the two previous cases, in this situation, little depends on you.