Question GPU overheating?

[Mr Helix]

Hi, a year ago I undervolted my 3070 MSI 2x, since it was overheating to 83 degree when playing basically anything, and idle at 50+
Lately I installed windows 11 and ofc had to reconfigure the undervolt, but this time it didn't work at all.
The temps just don't drop anymore and
After changing thermal paste, it's overheating in demanding games like rdr2 or dying light 2,etc.(83 and up) and also I went to furmark and what I noticed was VRAM overheating to 100+ degrees.

I read that thermal pad change should fix it so I ordered 2.25 and 2.75 since that's the recommended and right sizes sourced from official MSI website.

I'm gonna change pads, until then is there anything else i can do?? Thanks in advance!

My PC is well set up with fans, dust free, I have 5 fans 3 intake 2 outtake

 

[JohnBonhamsGhost]

I undervolted my 3070 MSI 2x, since it was overheating to 83 degree when playing basically anything, and idle at 50+...
changing thermal paste...gonna change pads

normally a GPU isn't reaching these temperatures by poor design, but it can happen.

you should always read reviews for the particular model to see if others report similar issue before attempting to manually correct such an issue.

usually it is the user's environment causing these types of temperatures.

is there anything else i can do?
My PC is well set up with fans, dust free, I have 5 fans 3 intake 2 outtake

the amount of fans alone isn't the only property in relation to controlling temperature.

what make & model case; how well ventilated,
what make & model fans; what is their CFM, what is their mm/H2O,
where are fans located, what is their orientation,
what type of fan speed curve profiles are being used,
what is the ambient temperature where the system is located,
what thermal paste is being used, how is it applied?

include the make & model of all system components.

 

[Mr Helix]

normally a GPU isn't reaching these temperatures by poor design, but it can happen.

you should always read reviews for the particular model to see if others report similar issue before attempting to manually correct such an issue.

usually it is the user's environment causing these types of temperatures.

the amount of fans alone isn't the only property in relation to controlling temperature.

what make & model case; how well ventilated,
what make & model fans; what is their CFM, what is their mm/H2O,
where are fans located, what is their orientation,
what type of fan speed curve profiles are being used,
what is the ambient temperature where the system is located,
what thermal paste is being used, how is it applied?

include the make & model of all system components.

This is my case, I don't have a front panel on for airflow since it's just solid glass .

I got 3 front intake and on back outtake and 1 one more outtake upwards.

 

[Mr Helix]

normally a GPU isn't reaching these temperatures by poor design, but it can happen.

you should always read reviews for the particular model to see if others report similar issue before attempting to manually correct such an issue.

Watching game tests and other people's 3070s max 60-70 during load on very demanding games, and idle on 20-30-40.
Whilst mine is 50 or more on idle and 80 to 90 in low end games cornerns me
And like I had same temps back in cold winter, it just maybe added 3 degrees now it's summer now.

 

[Phaaze88]

This is my case, I don't have a front panel on for airflow since it's just solid glass .

I got 3 front intake and on back outtake and 1 one more outtake upwards.

Well, that's one piece of the puzzle. The fans at the front can hardly get any air, if at all - doesn't matter if there's 3 of 'em there. Solid and semi-solid front panels are generally frowned upon because of this. If this is a style you like on a case, know that you tradeoff cooling performance over ALL of your devices for aesthetics.

3070 MSI 2x

Another piece, the Ventus 2X - or rather, the front and rear exhaust cooler that it uses. About half of the waste heat exits the front of the card, which fights with any front case intakes, and... well, it's heating up case ambient, plus some of that front exhaust is going below it to get sucked back in.
[Ventus 3X mostly exhausts at the sides, so the airflow is going to be a bit different.]

It's kinda heating itself up here. Combined with the poor case intake, the current situation seems to be about right.

I read that thermal pad change should fix it so I ordered 2.25 and 2.75 since that's the recommended and right sizes sourced from official MSI website.

No one ever talks about pad hardness(soft/firm/hard/etc), which matters just as much as pad depth. Hardness is going to be the difference between worse/improved memory thermals, and/or worse gpu thermals, or even a cracked gpu die. Hardness info for the original pads isn't disclosed, it seems, unless it's for a specific model(like Kritical's thermal pads, for example).



Try not to do too many things at one time. First thing I suggest trying is removing the front and side panels, then check gpu core, hot spot, and memory junction again in game.

 

[Mr Helix]

Well, that's one piece of the puzzle. The fans at the front can hardly get any air, if at all - doesn't matter if there's 3 of 'em there. Solid and semi-solid front panels are generally frowned upon because of this. If this is a style you like on a case, know that you tradeoff cooling performance over ALL of your devices for aesthetics.


Another piece, the Ventus 2X - or rather, the front and rear exhaust cooler that it uses. About half of the waste heat exits the front of the card, which fights with any front case intakes, and... well, it's heating up case ambient, plus some of that front exhaust is going below it to get sucked back in.
[Ventus 3X mostly exhausts at the sides, so the airflow is going to be a bit different.]

It's kinda heating itself up here. Combined with the poor case intake, the current situation seems to be about right.


No one ever talks about pad hardness(soft/firm/hard/etc), which matters just as much as pad depth. Hardness is going to be the difference between worse/improved memory thermals, and/or worse gpu thermals, or even a cracked gpu die. Hardness info for the original pads isn't disclosed, it seems, unless it's for a specific model(like Kritical's thermal pads, for example).



Try not to do too many things at one time. First thing I suggest trying is removing the front and side panels, then check gpu core, hot spot, and memory junction again in game.

Thanks for the input fam, as I said I don't use the front panel for airflow so it's always off.

And GPU is used, I bought it and it's been opened before and what I'm thinking is that whoever had it before had problem with overheating surely and he might have also switched the thermal pads, but he just used the wrong size, since people here aren't smart and the correct sizes can't be found here (yes people are dumb enough here to not know how to order stuff online)