Saturday
bonjour , j'ai un ROG Strix G18 G814 avec un Intel core i9-13980hx et une RTX 4080 je stocke que mon processeur chauffe beaucoup pour un pc neuf jusqu'à 90 degrés en jeu ( en perf max ) , du coup j'ai Undervolter mon cpu avec le logiciel XTU de Intel , j'ai baisser core tension offset, le processeur cache tension offset, et le efficace cores cache tension offset à -0.125 ( j'ai baisser de -0.025 à chaque fois et j'ai fais un Bench avec Cinebench R15 du coup je suis passer à 10 degrés de moins ,et sur Cinebench R15 , d'un score de 4978 cb à -0.025 à un score de 5358 cb à -0.125 c'est la bonne méthode ou pas pour l' Under volt ? c'est trop ? vu que sur se logiciel on pouvait faire un overclock automatique c'est conseiller ou pas ? risques ? merci pour vos raiponces
Solved! Go to Solution.
Sunday - last edited Sunday
These are your boost power limits and if you followed that youtube video and placed PL1, PL2 and IccMAX to unlimited this is why the CPU will run hotter when under load.
The default boost design places a power limit on the CPU so that the cores cannot all run at the maximum boost frequency at the same time. For example, by default when all your P-cores are fully loaded they do not run at 5.6GHz, they will generally be around 4GHz (maybe more / less). This is because of the limits in place / design spec, The CPU is only allowed to consume up to 157W Intel spec in boost mode.
If you change the PL1, PL2 and IccMAX to unlimited it means the CPU will try to use as much power as it can during boost mode, in Cinebench your CPU would likely have been running all cores higher than 4GHz, that means it would have been using a lot more power and consequently generating a lot more heat despite undervolting.
Also something to think about when undervolting. Your CPU is always trying to boost itself no matter what, if you make it run cooler it has the potential to boost higher, to the point that undervolting may have minimal impact.
In this test, -0.8 sees 4.3GHz vs 3.7GHz average across the P-Cores with minimal impact to load temperatures. The only improvement is the benchmark score as it was able to boost higher.
https://www.youtube.com/watch?v=w24rSJpUS6s
https://www.youtube.com/watch?v=fRRN0MUmUnk
For example, CPU boosts to 3.7GHz runs 100 DegC.
We undervolt it so now at 3.7GHz it runs cooler.
CPU knows this so will increase the boost until it hits 100 Deg C again.
So undervolting has helped it run cooler but the boost mode took advantage instead, so in practice it doesn't translate to that much lower temperatures when it is under a very high load.
And not matter what you do, if a single core boosts to 5.6GHz it will always show a higher temperature in games.
But before adjusting settings you need to determine how hot / how efficient your cooler is performing when the CPU is running base frequencies as if there is an issue it will be hot no matter what you do.
I have more testing and explanation here (give hardware monitor a go and try running some games too, like I show)
https://www.electrostingz.com/2025/06/how-to-check-laptop-cooler-performance-intel/
If your base frequency temperature is terrible, fix that first then start adjusting settings, I would expect your CPU to be about the same as my 14900HX at base frequencies. This will also show you how hot your laptop will run as a minimum too. (Under 75 Deg C is good in a non boost state.)
Saturday - last edited Saturday
Hello,
It depends on what it is you are trying to achieve but the temperatures at default settings are normal, the CPU will easily get over 90 Deg C.
Cinebench multicore testing is also not representative of a gaming environment as the CPU E-cores are never used or fully loaded, for the most part these are almost constantly idle
And also when gaming your RTX 4080 might use 175W, in this state the CPU only has 65W available so it will be limited in terms of performance.
I mention all of this as if you focus on the average CPU frequency during gaming it is not going to be 5.6GHz all cores but changing rapidly to the point that you could be much better off with limiting the CPU boost ratio instead if you want lower temperatures.
For example if you set the P-cores to x40 (or other) and test gaming performance / benchmarks / temperatures, how does it compare to default? (forget benchmarks that use E-cores).
Do not use the automatic overclock feature, its mainly for desktop processors.
Sunday
salut ,
merci pour ta raiponce , cependant je suis un amateur consternant l'overclocking et l'Under volt, j'ai suivit une vidéos, sur YouTube pour gagner uniquement en température (https://www.youtube.com/watch?v=HOnCZnokdnA), car je veux éviter que mon processeur chauffe prématurément , je joue principalement a des RPG , et des jeux de survie , en mode solo je cherche avant tout un rendu agréable, et que mon ordinateur ne soit pas un four , merci pour ton conseil sur l'overclock automatique ,tu maitrise bien le sujet , j'ai armory crate et j'aurai voulu savoir a quoi ca ca cetr ( voir image ) et si je pouvait baisser la temperature avec ca ou faut pas y toucher
Sunday
en jeu
en jeu
Sunday - last edited Sunday
These are your boost power limits and if you followed that youtube video and placed PL1, PL2 and IccMAX to unlimited this is why the CPU will run hotter when under load.
The default boost design places a power limit on the CPU so that the cores cannot all run at the maximum boost frequency at the same time. For example, by default when all your P-cores are fully loaded they do not run at 5.6GHz, they will generally be around 4GHz (maybe more / less). This is because of the limits in place / design spec, The CPU is only allowed to consume up to 157W Intel spec in boost mode.
If you change the PL1, PL2 and IccMAX to unlimited it means the CPU will try to use as much power as it can during boost mode, in Cinebench your CPU would likely have been running all cores higher than 4GHz, that means it would have been using a lot more power and consequently generating a lot more heat despite undervolting.
Also something to think about when undervolting. Your CPU is always trying to boost itself no matter what, if you make it run cooler it has the potential to boost higher, to the point that undervolting may have minimal impact.
In this test, -0.8 sees 4.3GHz vs 3.7GHz average across the P-Cores with minimal impact to load temperatures. The only improvement is the benchmark score as it was able to boost higher.
https://www.youtube.com/watch?v=w24rSJpUS6s
https://www.youtube.com/watch?v=fRRN0MUmUnk
For example, CPU boosts to 3.7GHz runs 100 DegC.
We undervolt it so now at 3.7GHz it runs cooler.
CPU knows this so will increase the boost until it hits 100 Deg C again.
So undervolting has helped it run cooler but the boost mode took advantage instead, so in practice it doesn't translate to that much lower temperatures when it is under a very high load.
And not matter what you do, if a single core boosts to 5.6GHz it will always show a higher temperature in games.
But before adjusting settings you need to determine how hot / how efficient your cooler is performing when the CPU is running base frequencies as if there is an issue it will be hot no matter what you do.
I have more testing and explanation here (give hardware monitor a go and try running some games too, like I show)
https://www.electrostingz.com/2025/06/how-to-check-laptop-cooler-performance-intel/
If your base frequency temperature is terrible, fix that first then start adjusting settings, I would expect your CPU to be about the same as my 14900HX at base frequencies. This will also show you how hot your laptop will run as a minimum too. (Under 75 Deg C is good in a non boost state.)
Monday
merci beaucoup , https://www.electrostingz.com/2025/06/how-to-check-laptop-cooler-performance-intel/ ,m'a bien aider , en se qui consterne la ventilation je règle ca sur Gheller et je possède 3 ventilateur , j'aimerais savoir si il est préférable d'augmenter mes ventilateur en mode turbo au max ou laisser d''usine ( voir photos) le bruit me dérange pas , je joue avec un casque .MODE USINE
Max
Monday
I find Default Turbo fan profiles to be ok, max fans are too loud 🙂