How can I limit my CPU temps?

If you notice any weird behavior such as freezes, system hang ups, restarts, blue screen, then you get the point. If nothing happens, then it is successful. Check your CPU temperature. Is it lower now? How much lower?

spapakons said:

If you notice any weird behavior such as freezes, system hang ups, restarts, blue screen, then you get the point. If nothing happens, then it is successful. Check your CPU temperature. Is it lower now? How much lower?

Click to expand...

alright, thank you i got that! does 100 seem like a okay amount to undervolt it by tho

and yes, it has been way lower without taking away too much performence

Yes, 100 mv is about average to a little high undervolt, so if it is stable you have done well. Congratulations!

What usually happens when you have gone too far is that you start getting recoverable WHEA errors (and eventually a unrecoverable one with a blue screen windows crash). But the OCCT stress program checks for WHEA recoverable errors and reports them.

You can also see if you have had any recoverable WHEA errors in one of the system log files. I'll try to remember to post where later today.

Cheers

PerpetualCycle said:

Yes, 100 mv is about average to a little high undervolt, so if it is stable you have done well. Congratulations!

What usually happens when you have gone too far is that you start getting recoverable WHEA errors (and eventually a unrecoverable one with a blue screen windows crash). But the OCCT stress program checks for recoverable errors and reports them.

You can also see if you have had any recoverable WHEA errors in one of the system log files. I'll try to remember to post where later today.

Cheers

Click to expand...

noted, ill try to do more long OCCT tests and just generally do my daily activites to test it. assuming i get a BSOD, what do you think i should reduce it to?

also thank you! im not sure if ill have any yet because i just recently undervolted it

If you want to stress your computer to check stability, you can use a utility such as Passmark BurnInTest. It checks CPU, RAM, graphics, sound and hard disk performance. The trial version has a fixed duration of 15 minutes but it should be enough. It also monitors temperatures and warns you if they exceed a certain threshold.

zqchri said:

noted, ill try to do more long OCCT tests and just generally do my daily activites to test it. assuming i get a BSOD, what do you think i should reduce it to?

also thank you! im not sure if ill have any yet because i just recently undervolted it

Click to expand...

10 mv at a time.

PerpetualCycle said:

10 mv at a time.

Click to expand...

Just don't overdo the testing with intensive stress tests to avoid wearing your processor. It would be good to make regular system backups with a program like Macrium Reflect.

PerpetualCycle said:

10 mv at a time.

Click to expand...

mv?

PerpetualCycle said:

Just don't overdo the testing with intensive stress tests to avoid wearing your processor. It would be good to make regular system backups with a program like Macrium Reflect.

Click to expand...

noted, thank you! ill just do what i usually do to test the undervolt, with one or two stress tests.

zqchri said:

mv

Click to expand...

1/1000 of a volt

PerpetualCycle said:

1/1000 of a volt

Click to expand...

ou damn so like i go down to 99.999?

zqchri said:

ou damn so like i go down to 99.999?

Click to expand...

No 90. The 100 is also mv units.

zqchri said:

how would i know if the undervolt doesnt turn out good?

Click to expand...

You could always decide to run a stress test for a long period while the CPU is undervolted by 100mV (mV = milli-volts so, 1mV = one thousandth of a volt) or run a stress test when you aren't going to use the laptop for at least another while. So, each time when you need to use the latptop, you can always just end the stress test and undo the undervolting, i.e. until you are ready to re-apply the undervolting and let the laptop begin a new test session. The bigger the total amount of time that the test will have been able to run on the undervolted CPU without causing any stability issues to be revealed by it so far, the more likely it will have become that it's "good". However, if it's already been running stable like this for a whole week, then this still doesn't guarantee anything besides the fact that it looks stable. In addition, a lot of people will agree that letting the test run for a whole week in total is not exactly very practical. That's why, if it still looks stable after letting it run overnight, you could then decide to undervolt it by an additional 10mV, and test it like that for another night. Keep repeating this until it no longer looks stable. Then put it back to what it was when it still looked stable (or maybe consider to back off a little bit farther, like, just to be a little more on the safe side).

hdmi said:

You could always decide to run a stress test for a long period while the CPU is undervolted by 100mV (mV = milli-volts so, 1mV = one thousandth of a volt) or run a stress test when you aren't going to use the laptop for at least another while. So, each time when you need to use the latptop, you can always just end the stress test and undo the undervolting, i.e. until you are ready to re-apply the undervolting and let the laptop begin a new test session. The bigger the total amount of time that the test will have been able to run on the undervolted CPU without causing any stability issues to be revealed by it so far, the more likely it will have become that it's "good". However, if it's already been running stable like this for a whole week, then this still doesn't guarantee anything besides the fact that it looks stable. In addition, a lot of people will agree that letting the test run for a whole week in total is not exactly very practical. That's why, if it still looks stable after letting it run overnight, you could then decide to undervolt it by an additional 10mV, and test it like that for another night. Keep repeating this until it no longer looks stable. Then put it back to what it was when it still looked stable (or maybe consider to back off a little bit farther, like, just to be a little more on the safe side).

Click to expand...

i dont think i uavw that much time on my hands and im not allowdd to leave my laptop running overnight, so do you think i should just leave it be and when i get errors i see then?

PerpetualCycle said:

No 90. The 100 is also mv units.

Click to expand...

ah alright !

zqchri said:

hi hi,

im just posting this because im a little confused. with my laptop build, i always have to end up using the lowest graphics on any game and have only that game open to be able to actually keep at a temp of 7-80, and I find that sort of annoying as I want to be able to use other apps, such as discord or google, without having to worry about my laptop overheating. is there any way i can do this?

to note, i understand that my GPU, the intel core i9, does get hot easily. and adding on, im on a laptop as well. i usually played plugged in as well. would that be the problem, just playing while plugged in? i do have a cooling pad as well. im just quite unsure what to do here, any help is useful!

Click to expand...

I don't know why people choose laptops for gaming, every single laptop I have ever used heat always becomes an issue. Laptops are not designed to dissipate heat efficiently and most laptops are not built for upgrading or adding better hardware. Especially were gaming is involved that is more graphics and processor intensive , when gaming, both of these generate tons of heat.

My advice? ...... use your laptop for normal use, move your gaming to a desktop with a tower that has some room inside it where you can manage the heat from your processor and graphics more efficiently, then you will have a lot more cooling options if needed. Cooling pads work on laptops but only to a small extent , someone made millions off of something that really does not do all that much.
If you do not have good airflow through the laptop case the cooling pad becomes null, because all you are really doing is just cooling down the plastic encasement, you're not moving the heat out, I have yet to see a laptop designed that has good ventilation.
If you cannot push the hot air out and reduce the ambient temperature inside the laptop case by pushing cool air through , you are going to battle processor and graphics card temperatures until the cows come home or your processor and/or graphics card dies, whichever comes first.

I never recommend laptops for gaming , unless you are just a gaming junkie that just has to have their games wherever they go and in that case I tell them ..... roll the dice.

Like water, heat and electronics do not play well together.

i get your point, but at the point where i got this laptop i really did not have any space for a pc

zqchri said:

It definately goes down a LOT when not plugged in, but it barely lasts

Click to expand...

Because unplugged is power throttling (less power = lower voltages/frequencies = lower temperatures).

Fabler2 said:

I just run in Balanced plan and set Advanced settings 'Max processor state' to 99 or 98% to stop it going to 90℃ Runs at 74℃ in this state. It does run at slightly less performance but that's the trade-off with a laptop.

View attachment 105735

Click to expand...

Setting the Maximum Processor State to 99% - is basically "Disabling Turbo Boost" - thus, the CPU will run at 2.2 GHz - where the Maximum Turbo Boost for his CPU goes up to 5.5 GHz. To be fair... high temperatures will trigger throttling - so it rarely stays at Max "on a laptop". But still, that's a pretty high limit to set for that CPU - even tho - obviously, since the CPU is running at less than 50% of its Max potential - the temperatures are significantly lower as well.

That being said - i do recommend using a power profile with Turbo Boost disabled - for general use - even some games (it's a powerful modern CPU - so even at half its potential - can still do a good enough job in some games, but definitely more than good enough for general use). Also recommend using:


- to easily switch between Power Profiles depending on needs, but also for its main goodies - since this is one of the best tools for Undervolting and power throttling management. The creator of ThrottleStop is really helpful too:


Tho, in case its performance is rather low for certain games with Turbo Boost disabled - i recommend creating another power profile - where the CPU is limited to around 4 Gz. You'll have to run some tests - see what's the max frequency at which your laptops will reach reasonable Max temperatures. Some coolers are slightly (not much room for improvement - on such tight space, as in - very limited airflow and coolers using small fans too) better - being able to keep the temperatures below 90*C even at 4.8 Ghz - while others struggle even with 4 Gz - in which case 3.8 gz can be more suitable.

Your GPU needs Undervolting as well - and for this you can actually use a tool made for MSI: Afterburner Also, an easy guide:

Such Laptops come with various options in the manufacturers App. Mine has what it calls Thermal modes and GPU Working Modes. No doubt MSI supplies similar sort of thing.

That is the first thing to look at before any additional "Managers".

Not knowing what mV is from kV, or µV, stroll on.

neves said:

Your GPU needs Undervolting as well - and for this you can actually use a tool made for MSI: Afterburner Also, an easy guide:

Click to expand...

my gpu temps usually seem fine, are you sure this is needed?