nvidia ada lovelace/RTX 4000 series. safe hot spot temp

[Tomgang]

So a quick question.

I just got my hands on gigabytes new little RTX 4060 with low profile cooler. So cooling is off cause limited.

Power is locket to the stock 115 watt. So power cant be raised.

Now i am playing around with overclock to get as much out of it as possible. So 200+ on GPU or 2880 mhz to 2900 mhz with a lillte undervolt from stock 1.05 mv to 1.025 mv and right now 1700+ on memory in msi after burner. Thats deffently gives a little bump in performance.

But here comes my possible concern. specially hot spot temp. right not gpu temp hovering around 76 C to 80 C peak. While hot spot hovering around 92 C to 95 C with a short peak of 97 C.

I find it a little toasty for the hot spot. But remember it´s a low profile cooler on it with 3 fans, so cooling is limited. It´s more are this safe hot spot temp (the normal llgpu temp are high, but should be safe, my concerns is mostly the hot spot) or should i dial back to lower hot spot temp?

 

[Deleted member 182555]

try keeping case open, run 3dmark timespy and put gpu score... I got 12731 points with biosmod wip..

 

[Tomgang]

try keeping case open, run 3dmark timespy and put gpu score... I got 12731 points with biosmod wip..

That will first be tomorrow as im about to hit that sack. But i cant see why that would change temp.

The temperature i got is from normal gaming while testing overclock stability in an older game (Dying light dlc the following). Overclock seems stable, but i am concerned about hot spot temp.

 

[Deleted member 182555]

I have currently removed rtx 4060 from the PC, because I'm experimenting on other GPUs.
I can't tell you anything about my temps, I haven't marked them. check real core speed with rtss you need to add it to the overlay.
with the low profile GPU you will always have to make trade-offs for heat.

 

[Tomgang]

I have currently removed rtx 4060 from the PC, because I'm experimenting on other GPUs.
I can't tell you anything about my temps, I haven't marked them. check real core speed with rtss you need to add it to the overlay.

I can do a test tomorrow as stated. But mine and your score might differ to much.

Reasons for that is your cpu is older and running PCIe gen 3. That might be a bottleneck in it self as this silly card only runs PCIe x8. Also what power taget dit you do the run at. Stock or above the 100 %. My card is locket to 100 % or Stock 115 watt. So if you ran with a higher power target, that will definitely also make our comparison differ.

 

[Deleted member 182555]

TDP had been modded to 170W, but I lowered it a bit because it was excessive, it's risky...

I'm curious to see how the lowprofile goes.

 

[Tomgang]

TDP had been modded to 170W, but I lowered it a bit because it was excessive, it's risky...

I'm curious to see how the lowprofile goes.

170 watts is deffently high for a rtx 4060. You should be safe with 8 pin + PCIe power when it comes to that as that is 150 watt + 75 watt for PCIe power. But Yeah thats risky when it comes to vrm temp and life span. You might risk roasting something on the card with 170 watt. Rtx is not designed for 170 watt, not even the overbuild Asus strix version.

Actually i am curious to. I will see if i can get a run done after i come home from work.

 

[Vya Domus]

Seems fine, I mean it's a miniscule cooler what do you expect.

 

[Deleted member 182555]

170 watts is deffently high for a rtx 4060. You should be safe with 8 pin + PCIe power when it comes to that as that is 150 watt + 75 watt for PCIe power. But Yeah thats risky when it comes to vrm temp and life span. You might risk roasting something on the card with 170 watt. Rtx is not designed for 170 watt, not even the overbuild Asus strix version.

Actually i am curious to. I will see if i can get a run done after i come home from work.

I should have the video somewhere if I haven't deleted it, from that you can see gpu temps.

 

[Tomgang]

I should have the video somewhere if I haven't deleted it, from that you can see gpu temps.

So i made a run. Power limited big time because of the stock power limit. Keep GPU score down. PEak hot spot temp was 93 C

Seems fine, I mean it's a miniscule cooler what do you expect.

I never exspected great temp from it. I only wanted to know if ot was with in safe limits. To high temp will degrade or even damage the gpu in the long run if to high.




So to all. I would still like to know if these high hot spot temp is safe or not?

 

[R-T-B]

So to all. I would still like to know if these high hot spot temp is safe or not?

Card will generally throttle to keep the temps safe and in line, so I don't think it's going to be much of an issue. At worst you'd lose some performance due to throttling.

 

[Tomgang]

Card will generally throttle to keep the temps safe and in line, so I don't think it's going to be much of an issue. At worst you'd lose some performance due to throttling.

Alright. I am aware of thermal throttle. Thanks for your reply.

 

[Deleted member 182555]

So i made a run. Power limited big time because of the stock power limit. Keep GPU score down. PEak hot spot temp was 93 C

View attachment 314328


I never exspected great temp from it. I only wanted to know if ot was with in safe limits. To high temp will degrade or even damage the gpu in the long run if to high.




So to all. I would still like to know if these high hot spot temp is safe or not?

a little low but I expected it, the Asus 4060 dual does 12200pts@138w with normal MSI afterburner overclocking.
if you have some space in the case you can try installing a fan that blows hot air away from that area. for undervolt offset use multiples of 15mhz...

 

[Tomgang]

a little low but I expected it, the Asus 4060 dual does 12200pts@138w with normal MSI afterburner overclocking.
if you have some space in the case you can try installing a fan that blows hot air away from that area. for undervolt offset use multiples of 15mhz...

Yeah as stated, i am very power limited.

I have a lot of space in my case. Phanteks enthoo 719. I am actually right this moment talking with a person about some 3d printed stuff to direct more airflow to the system. Airflow right now is not great.

 

[Beginner Macro Device]

A little bit worrying temperature. Decrease your wattage the way your hotspot stays below 88C and you'll be completely fine.
I don't say 97 is dangerous. It's just a little concerning. At 88C, your GPU will live basically forever.

 

[BMfan80]

By
Tose Nikolov
Computer programmer that works with embedded systems

This is a complicated topic.

Basically, on the GPU die there are hundreds of temperature sensors. What GPU hotspot temperature shows you, is the maximum measured temperature of all the sensors.

The temperature of concern on a silicon chip is 125C. You don’t want any part of your chip to be above 125C, because at that temperature the structure of the chip starts degrading.

In ye olden days, all you had was 1 temperature sensor on the middle of the chip, and then the GPU engineers would calculate the reported temperature, when parts of the chip got toasty. So they would calculate that when there are hot spots on the chip of 120C(5C gap for safety), the temperature sensor would show 95C. So they would set the max safe temperature of the GPU at 95C. And as long as you were below 95C, you would be pretty sure that no part of your GPU was overheating.

Nowadays, you get the sensor grid, so now the GPU manufacturers get 2 numbers to play with. They get the “average” measured temperature, and the “hotspot” measured temperature. This temperature measurement still happens on the surface of the silicon chip, so there are still hotspots inside the chip that are hotter that the “GPU hotspot temperature”. The process is the same, GPU designers calculate the max safe temperature of the GPU chip, usually 120C again, and then see what numbers they are getting on the “average temperature” and “hotspot temperature” sensor. Usually the average temperature sensor is in the range of 92–97C depending on GPU. The safe temperature for the “hotspot temperature” sensor, is in the 110C range. Basically, the “hotspot temperature” sensor is closer to the true value you care about, but still only halfway there.

So. As long as the GPU temperature is below 95C, and the hotspot temperature is bellow 110C, you will be fine.

 

[iameatingjam]

I find it a little toasty for the hot spot. But remember it´s a low profile cooler on it with 3 fans, so cooling is limited. It´s more are this safe hot spot temp (the normal llgpu temp are high, but should be safe, my concerns is mostly the hot spot) or should i dial back to lower hot spot temp?

Depends... I mean, was that after 5 hours of intensive gaming? Then I wouldn't be too concerned. Is that happening right after you fire up a game or benchmark? in that case its.. not good. But its probably managable. I had that problem with my 3070 ( it was both hot and had a very wide delta). I was able to somewhat solve this by setting the fan to 80% minimum, undervolting it and replacing the thermal paste and pads.... (multiple times.... long story...) Anyway, make sure you change that fan curve, (they're small but... you can still use them) and undervolt as much as you can. Its pretty safe so go slow but see how low you can go with things remaining stable.

And when it comes to the overclocking, I mean if you must, overclocking the memory is going to have less affect on temperature than the gpu itself. Maybe focus on that more? EDIT: I don't necessarily mean go higher than your 1700 thats already pretty high but you might find that its doing more for your performance than the core overclock, depends on the card and the game.

I never exspected great temp from it. I only wanted to know if ot was with in safe limits. To high temp will degrade or even damage the gpu in the long run if to high.

Well from what I understand the silicon doesn't start to rapidly degrade until we get into 120-130s. But will it shorten the lifetime of the card? Possible. Also you can't always trust that these readings are 100% accurate so its pretty important to leave some buffer.

Do you think you're going to have this card for a long time?

 

[Tomgang]

A little bit worrying temperature. Decrease your wattage the way your hotspot stays below 88C and you'll be completely fine.
I don't say 97 is dangerous. It's just a little concerning. At 88C, your GPU will live basically forever.

I have all ready lowered wattage by undervolting a little.

By
Tose Nikolov
Computer programmer that works with embedded systems

This is a complicated topic.

Basically, on the GPU die there are hundreds of temperature sensors. What GPU hotspot temperature shows you, is the maximum measured temperature of all the sensors.

The temperature of concern on a silicon chip is 125C. You don’t want any part of your chip to be above 125C, because at that temperature the structure of the chip starts degrading.

In ye olden days, all you had was 1 temperature sensor on the middle of the chip, and then the GPU engineers would calculate the reported temperature, when parts of the chip got toasty. So they would calculate that when there are hot spots on the chip of 120C(5C gap for safety), the temperature sensor would show 95C. So they would set the max safe temperature of the GPU at 95C. And as long as you were below 95C, you would be pretty sure that no part of your GPU was overheating.

Nowadays, you get the sensor grid, so now the GPU manufacturers get 2 numbers to play with. They get the “average” measured temperature, and the “hotspot” measured temperature. This temperature measurement still happens on the surface of the silicon chip, so there are still hotspots inside the chip that are hotter that the “GPU hotspot temperature”. The process is the same, GPU designers calculate the max safe temperature of the GPU chip, usually 120C again, and then see what numbers they are getting on the “average temperature” and “hotspot temperature” sensor. Usually the average temperature sensor is in the range of 92–97C depending on GPU. The safe temperature for the “hotspot temperature” sensor, is in the 110C range. Basically, the “hotspot temperature” sensor is closer to the true value you care about, but still only halfway there.

So. As long as the GPU temperature is below 95C, and the hotspot temperature is bellow 110C, you will be fine.

Thanks, very useful. So bacically i should be safe accordingto this.

Depends... I mean, was that after 5 hours of intensive gaming? Then I wouldn't be too concerned. Is that happening right after you fire up a game or benchmark? in that case its.. not good. But its probably managable. I had that problem with my 3070 ( it was both hot and had a very wide delta). I was able to somewhat solve this by setting the fan to 80% minimum, undervolting it and replacing the thermal paste and pads.... (multiple times.... long story...) Anyway, make sure you change that fan curve, (they're small but... you can still use them) and undervolt as much as you can. Its pretty safe so go slow but see how low you can go with things remaining stable.

And when it comes to the overclocking, I mean if you must, overclocking the memory is going to have less affect on temperature than the gpu itself. Maybe focus on that more? EDIT: I don't necessarily mean go higher than your 1700 thats already pretty high but you might find that its doing more for your performance than the core overclock, depends on the card and the game.


Well from what I understand the silicon doesn't start to rapidly degrade until we get into 120-130s. But will it shorten the lifetime of the card? Possible. Also you can't always trust that these readings are 100% accurate so its pretty important to leave some buffer.

Do you think you're going to have this card for a long time?

In short: yes, potentially i will have this card for a few years at least. Cause it´s not often new and good low profile cards comes out.