Jump to content

How do gpu vrm's work

remo233

Ok so I recently bought an rtx 4070 super. I had a question about what determines how hot the vrms get. Vrm temps in game;

 

Forza horizon 5 4k max settings -vrm temp 78

Far cry 6 4k max settings- vrm temp 84

Resdient evil 3 4k max settings- vrm temp 92

Dead spake remake 4k dlss quality max settings- vrm temp 86

Palwolrd 4k high settings- vrm temp 84

Cyberpunk 4k performance dlss overdirve rt- vrm temp 80

 

The outliner here is resident evil.

Link to comment
Share on other sites

Link to post
Share on other sites

It will depend on two factors: How much power is running through them (i.e. essentially how much GPU load is there?), and how much cooling do they get.

 

If they run under high load, but at the same time the cooler is running slower because the GPU core isn't getting as hot, that might result in them getting hotter, despite power draw being the same.

Remember to either quote or @mention others, so they are notified of your reply

Link to comment
Share on other sites

Link to post
Share on other sites

3 minutes ago, remo233 said:

Ok so I recently bought an rtx 4070 super. I had a question about what determines how hot the vrms get. Vrm temps in game;

 

Forza horizon 5 4k max settings -vrm temp 78

Far cry 6 4k max settings- vrm temp 84

Resdient evil 3 4k max settings- vrm temp 92

Dead spake remake 4k dlss quality max settings- vrm temp 86

Palwolrd 4k high settings- vrm temp 84

Cyberpunk 4k performance dlss overdirve rt- vrm temp 80

 

The outliner here is resident evil.

 

1 minute ago, Eigenvektor said:

It will depend on two factors: How much power is running through them (i.e. essentially how much GPU load is there?), and how much cooling do they get.

 

If they run under high load, but at the same time the cooler is running slower because the GPU core isn't getting as hot, that might result in them getting hotter, despite power draw being the same.

 Jumping on @Eigenvektors bandwagon here, what are the rest of the gpu temps youre seeing; gpu temp and gpu hotspot temp?

System specs:

 

 

CPU: Ryzen 7 7800X3D [-30 PBO all core]

GPU: Sapphire AMD Radeon RX 7800 XT NITRO+ [1050mV, 2.8GHz core, 2.6Ghz mem]

Motherboard: MSI MAG B650 TOMAHAWK WIFI

RAM: G.Skill Trident Z5 NEO RGB 32GB 6000MHz CL32 DDR5

Storage: 2TB SN850X, 1TB SN850 w/ heatsink, 500GB P5 Plus (OS Storage)

Case: 5000D AIRFLOW

Cooler: Thermalright Frost Commander 140

PSU: Corsair RM850e

 

PCPartPicker List: https://uk.pcpartpicker.com/list/QYLBh3

Link to comment
Share on other sites

Link to post
Share on other sites

7 minutes ago, TatamiMatt said:

 

 Jumping on @Eigenvektors bandwagon here, what are the rest of the gpu temps youre seeing; gpu temp and gpu hotspot temp?

Gpu temp itself does not go above 73 degrees. Hot spot in the high 80's can peak in the 90's but usually stays in the high 80',s.

Link to comment
Share on other sites

Link to post
Share on other sites

8 minutes ago, Eigenvektor said:

It will depend on two factors: How much power is running through them (i.e. essentially how much GPU load is there?), and how much cooling do they get.

 

If they run under high load, but at the same time the cooler is running slower because the GPU core isn't getting as hot, that might result in them getting hotter, despite power draw being the same.

Well in all games gpu usage is 100 percent or close to 100 percent.

Link to comment
Share on other sites

Link to post
Share on other sites

13 minutes ago, remo233 said:

Well in all games gpu usage is 100 percent or close to 100 percent.

Sure, that's how the GPU is designed to operate. But 100% usage doesn't really mean 100% of the GPU's resources are used at their fullest.

 

It simply means that it wouldn't be able to process commands any faster than it currently does. But it doesn't mean the commands it does get use all parts of the GPU core. So power draw and heat output can vary, even when the reported load is the same.

 

For example a game that is using raytracing or DLSS is going to make use of other parts of the GPU core than a game that isn't. And these different parts may draw a different amount of power.

Remember to either quote or @mention others, so they are notified of your reply

Link to comment
Share on other sites

Link to post
Share on other sites

8 minutes ago, remo233 said:

Gpu temp itself does not go above 73 degrees. Hot spot in the high 80's can peak in the 90's but usually stays in the high 80',s.

Moreso on a per game basis as youve detailed above

System specs:

 

 

CPU: Ryzen 7 7800X3D [-30 PBO all core]

GPU: Sapphire AMD Radeon RX 7800 XT NITRO+ [1050mV, 2.8GHz core, 2.6Ghz mem]

Motherboard: MSI MAG B650 TOMAHAWK WIFI

RAM: G.Skill Trident Z5 NEO RGB 32GB 6000MHz CL32 DDR5

Storage: 2TB SN850X, 1TB SN850 w/ heatsink, 500GB P5 Plus (OS Storage)

Case: 5000D AIRFLOW

Cooler: Thermalright Frost Commander 140

PSU: Corsair RM850e

 

PCPartPicker List: https://uk.pcpartpicker.com/list/QYLBh3

Link to comment
Share on other sites

Link to post
Share on other sites

26 minutes ago, remo233 said:

Gpu temp itself does not go above 73 degrees. Hot spot in the high 80's can peak in the 90's but usually stays in the high 80',s.

What's the brand and actual model of your 4070s? GPU temps vs hotspot temp, on average have 10-15c difference. Meaning if the gpu temp is around 65c, then hotspot is 75-80c. If the cooling solution and thermal pads are good, it could be lower. While vrm temp should be around 5-10c higher than gpu temp on average. However, if the differences are higher than those. You might want to check your thermal pads.

Personally, what I would do is first install afterburner(if you haven't yet). Then regardless of what gpu or hotspot temp is, I will set fan speed to manual and at around 70-80% fan speed(even 100%). If the vram is cooling down, it doesn't mean that the issue is already resolved. It just means that the fan curve is following the gpu temp/hotspot. Since those two aren't as hot, the fans are also not ramping up. But it will still just a band-aid solution. I would suggest googling to see if there are other people with the same make and model having thermal pad issue, because that might explain what's going on with your gpu. You can also check how much power is being feed to the gpu and try undervolting. 

Link to comment
Share on other sites

Link to post
Share on other sites

You can load VRAM to max and still have very low temps everything depends how hard is pushed memory controller load that is very depends from game but push to 100% basically impossible even stress test can't do that. Memory controller load will be higher if you turn on ray tracing or path tracing. Btw do undervolt 4xxx series run stupidly high clocks that generate massive heats by default settings and you gain 2-3% better performance only depends on luck you can run 15-20C cooler save a lot power and almost no performance hit.  I undervolt mine 4070 Ti from 1100mV 2790 Mhz to 935mV 2610 Mhz and I basically have no performance lose average 2 FPS but now GPU run average 15C less and use ~80W less

Link to comment
Share on other sites

Link to post
Share on other sites

15 hours ago, kitnoman said:

What's the brand and actual model of your 4070s? GPU temps vs hotspot temp, on average have 10-15c difference. Meaning if the gpu temp is around 65c, then hotspot is 75-80c. If the cooling solution and thermal pads are good, it could be lower. While vrm temp should be around 5-10c higher than gpu temp on average. However, if the differences are higher than those. You might want to check your thermal pads.

Personally, what I would do is first install afterburner(if you haven't yet). Then regardless of what gpu or hotspot temp is, I will set fan speed to manual and at around 70-80% fan speed(even 100%). If the vram is cooling down, it doesn't mean that the issue is already resolved. It just means that the fan curve is following the gpu temp/hotspot. Since those two aren't as hot, the fans are also not ramping up. But it will still just a band-aid solution. I would suggest googling to see if there are other people with the same make and model having thermal pad issue, because that might explain what's going on with your gpu. You can also check how much power is being feed to the gpu and try undervolting. 

It is an asus dual oc card. I have overclocked it. Reviews do talk about vrm temps on this caed being high but within safe operating parameters. Also, the parametersnyou have mentioned, the temps are not far off that except for one game, resident evil. I have not been too concernef about temps as I owned a gtx 1070 before that ran at 83 degrees with a hot spot in the 90'S. After this I had a founders edition rtx 3060 ti which ran between 78 and 80 degrees with a hot spot in the 90's. My only concern my residnlent evil made vram temps higher compared to many other games I have, even games like cyberpunk that I consider to be more demanding.

Link to comment
Share on other sites

Link to post
Share on other sites

11 hours ago, Winterlight said:

You can load VRAM to max and still have very low temps everything depends how hard is pushed memory controller load that is very depends from game but push to 100% basically impossible even stress test can't do that. Memory controller load will be higher if you turn on ray tracing or path tracing. Btw do undervolt 4xxx series run stupidly high clocks that generate massive heats by default settings and you gain 2-3% better performance only depends on luck you can run 15-20C cooler save a lot power and almost no performance hit.  I undervolt mine 4070 Ti from 1100mV 2790 Mhz to 935mV 2610 Mhz and I basically have no performance lose average 2 FPS but now GPU run average 15C less and use ~80W less

Thanks. Can I do the same for my 4070 super?

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×