cancel
Showing results for 
Search instead for 
Did you mean: 

About thermals on G512LW-WS

Undanup
Level 7
I have recently purchased this 2020 model and reaching temperatures up to 96°C por CPU and 87°C for GPU during gaming in turbo mode. It stabilizes between 92-95°C for CPU and 85-87°C por GPU, I have no flow obstructions and room temp is 20-25°C, it doesn't get better when I elevate the laptop . According to ASUS i should be getting better thermals since all 2020 models pack liquid metal cooling. In silent profile it can stay between 85-87°C for CPU and 74-78 for GPU, but this mode has a big impact on performance.
I have the latest BIOS, all windows and driver updates. I have been reaching to technical support in my ASUS account, but at this moment they are reccomending to reinstall windows, before I do that I would like to know if someone here has more information on wheter this is normal for this laptop model, or if you think a fresh Windows install will improve thermals.
Thanks in advance.
10,358 Views
28 REPLIES 28

Desani
Level 8
This is going to be a long post because I have just come up against this issue myself and have explored many options to control it.

I have a 3 week old G732LWS with the following tech specs:
17' Screen
Intel Core i7 - 10875h
Nvidia RTX 2070 Super

I will attempt to explain what is happening with thermals and why I doubt a re-install of the OS would do anything.

Like you when I first received my laptop, I noticed higher than what I would consider normal CPU temps and I was concerned, while running benchmarks, I was seeing temps hit as spike as high as 99°C and hovered around 89°C to 95°C. My GPU temps were a bit lower, hovering around the 75°C mark or lower. I was almost ready to issue a return at this point because the laptop is supposed to have superior cooling.

I explored online to see if this was something that was normal or if there was an issue with the laptop I received. I came to the conclusion that for these specs, having CPU temps in the 90's more or less how hot they run.

I was not satisfied to game as normal knowing that my laptop was slowly cooking itself into an early death. I started looking into what alternative options could be done maintain a constant lower temperature while minimizing the performance impact.

There are two main things that you will see people recommend right away when you read threads about people trying to attain lower temps.

1. Re-paste the thermal compound on the CPU/CPU
2. Undervolt your CPU

For option 1, there is no way I was going to open up a brand new laptop and attempt to re-paste, especially when there are some people saying that wasn't even possible when using liquid CPU paste. This option was off the table.

For option 2, because of a recent issue found with Intel CPU's called plundervolt. ASUS and other manufacturers have disabled the ability to Undervolt CPU's in their BIOS settings from Intel's recommendation. So if you are on a newer BIOS, you will no longer have the option to Undervolt. Thankfully my laptop came with bios version 307 and I have prevented any update to the bios so far. This version locks the ability to Undervolt in software from windows but it does allow me to setup a 80 mV undervolt in the BIOS itself. I don't know if a newer BIOS will allow you to undervolt but I am not going to risk upgrading at this point. The good news is, if you have upgraded, you should be able to downgrade as long as you have the BIOS image provided by ASUS for a previous BIOS version. I am not able to find a download for 307, my current version anywhere.

So here I am with a 80 mV undervolt and it really didn't do anything to my thermal temperature. CPU temps were still spiking high and were constantly in the 90's.

I disabled Hyper-threading in the BIOS and that drastically reduced temperatures but also drastically reduced performance. The performance hit was to large for me to be able to keep it disabled. I had Intel's tuning utility installed to be able to run some benchmarks early and I decided to look at what options were available for me adjust to be able to limit temperature.

The only three settings I could modify were:
Turbo Boost Long Power Max
Turbo Boost Short Power Max
Turbo Time Limit

Everything else was locked with my current BIOS.

The simple overview of these settings are: When the CPU hits it's highest base clock, it is able to ramp up compute power by going into Turbo Boost mode. The values allow you to control and limit the amount of power the CPU is allowed to use while entering Turbo mode.

I adjusted this values and ran benchmarks for every adjustment to find out if it was making a large thermal difference and it was.

So what I suspect is happening is, our laptops start to Thermal throttle the CPU at around 80°C but because of how powerful these chips are, the more power they can use the hotter the CPU gets and even with fans at max, it has a hard time keeping the CPU under it's thermal limit.

I ended up with setting the following settings for the Turbo Boost Power Limits:
Turbo Boost Long Power Max: 35
Turbo Boost Short Power Max: 100
Turbo Time Limit: 56

So with this setup, when the CPU needs to kick it into overdrive: it ramps up the power to the CPU to 100 watts for 56 seconds, and then it maintains a turbo boost at 35 watts after 56 seconds. During the short boost period, if CPU temperature is already high, it will spike up even higher in the 90's because of the 100 watt short power max, but after the 56 seconds, the CPU is only allowed to use 35 watts, original value was 45 watts. I found that with 45 watts as the long term value, the CPU was able to draw more power and generate more heat than the fan was able to dissipate fast enough. It would keep temperatures in the 90's but I was more comfortable with it being lower.

This allows me to maintain a temperature of between 76°C and 80°C for any long term duration. It will occasionally spike higher but the lower wattage allowance keeps thermals in check with a minor sacrifice to overall CPU performance. I was originally using XTU, Intel's utility, for modifing these values but once it was modified my PC had issues going to sleep. I then setup Throttlestop with the same values and apply them at every OS boot.

Next Steps:
If I want to look at limiting other sources of heat to attempt to get some performance back from the CPU Turbo mode, I could look at the following things:

Undervolt on the GPU: MSI Afterburner allows you to set custom voltage curves and in theory will allow for an undervolt on the GPU. This would reduce thermals on the GPU with no sacrifice to performance and would also remove some heat from the CPU as they used a shared heatpipe. I am still exploring this option but I am not comfortable enough with knowing what to do with the voltage curve yet so I haven't set this up.

Get a custom BIOS to unlock all the CPU voltage features in windows applications. If you are able to get on a BIOS that allows you to enable the XTU interface, you can in theory undervolt the CPU even more which will allow for better thermals with no performance it. It will also allow you a more options for turbo power limits and a limit on a per core bases that should allow you to tune power output so that you can achieve the same results without as large of a performance hit.

At this point, even without exploring the next steps, I am happy with the performance of the laptop as well as the thermals. I can game for an extended time and not worry about the laptop cooking itself to death. I will always continue to explorer better options, but even if nothing else pans out, this is still a beast of a machine that can keep the CPU temps in the high 70's.

Edit: I would also like to mention that I am able to game in all three ASUS profiles without worrying about thermals.
Silent Mode: Limit to CPU and GPU power as well as overall fan speed. Temps in the high 70's for CPU and GPU
Performance Mode: No limit on CPU and GPU Power but a slight limit to fan speed. Worst temps of the three profiles, low 80's for CPU temp.
Turbo Mode: Slight overclock to the GPU and GPU Mem and high fan speed: Temps similar or lower to Silent Mode

So I have found it fine to game in anything non-intensive in silent mode and anything that requires some more power, switching it to turbo mode. The only mode that I would not use that much is performance mode because it does put a limit to the fan speed which hinders the laptops ability to shed off heat and it does get a bit hotter than the other modes but nothing like it was before the changes to CPU turbo power.
Asus G732LWS
Geforce GTX 2070 Super
Intel i7-10875h

KorosuKivy
Level 10
Undanup wrote:
I have recently purchased this 2020 model and reaching temperatures up to 96°C por CPU and 87°C for GPU during gaming in turbo mode. It stabilizes between 92-95°C for CPU and 85-87°C por GPU, I have no flow obstructions and room temp is 20-25°C, it doesn't get better when I elevate the laptop . According to ASUS i should be getting better thermals since all 2020 models pack liquid metal cooling. In silent profile it can stay between 85-87°C for CPU and 74-78 for GPU, but this mode has a big impact on performance.
I have the latest BIOS, all windows and driver updates. I have been reaching to technical support in my ASUS account, but at this moment they are reccomending to reinstall windows, before I do that I would like to know if someone here has more information on wheter this is normal for this laptop model, or if you think a fresh Windows install will improve thermals.
Thanks in advance.


So yea I touched on this topic about the thermal output of the G512LV on this thread here https://rog.asus.com/forum/showthread.php?119431-G512LV-Preventing-permenant-boost-mode-on-mains even went as far as showing temps while playing games and idling go have a quick read on that and give me a more detailed idea on what your experiencing.

Unless you know how to repaste liquid metal ignore that. its too easy to damage the laptop if you screw it up since its a new purchase just send it back to your purchaser for a replacement...

But yea a reinstallation of the OS isn't going to fix the problem.

If you claim that your ambient room temperature is less than 20-25c id recommend a cooling mat from amazon.

I also have the G732LWS, and I have this issue. There's no way this can be normal. I've reinstalled the OS since I've had it, it does absolutely nothing.

I'm going to look into the Intel tuning thing that Desani talked about, that's the most helpful bit I've seen yet.

I'd be interested in knowing how the 2080 SUPER version fairs, being that it has 4 fan exhausts.

craig_cabbage wrote:
I also have the G732LWS, and I have this issue. There's no way this can be normal. I've reinstalled the OS since I've had it, it does absolutely nothing.

I'm going to look into the Intel tuning thing that Desani talked about, that's the most helpful bit I've seen yet.

I'd be interested in knowing how the 2080 SUPER version fairs, being that it has 4 fan exhausts.


I would recommend attempting the adjustments with Throttlestop rather than Intel® Extreme Tuning Utility. Intel's utility is more user friendly but I found that it caused more system instability.

To replicate my settings in throttlestop you would need to do the following:
Run a quick benchmark with hardware info monitoring the CPU and GPU for an accurate before picture
Download and unzip throttlestop
Launch throttlestop and click the button Turn on at the bottom of the window. That activate throttlestop settings and turns off monitoring mode.
We are going to ignore most of the values as the BIOS does not allow you to modify some of the important values that other guides online tell you to change.
The only values that I changed are in TPL menu.
In the TPL menu the first two items are the long and short power max values. Change the long value from 45 to 35. Change the short value from 125 to 100. I left the turbo time limit alone. The value that is going to make the most difference is the long time value. I enabled clap on both of these settings, from what I have read clap tells the processor to attempt to keep the values below or at the value supplied.
Once the new values are inserted, click okay and with throttlestop turned on, the change is immediate.

Run another benchmark with the new settings and make sure it runs for at least a couple of minutes. You know notice that after the first spike that lasts 56 seconds, the temperature will slowly decrease over time and settle on a value that should be lower than before because of the wattage limit. On my laptop, that took the overall temperature from the high 80's and put it in the high 70's. It might vary from computer to computer so play with the values that keep the temperature to a value you are okay with. Once you are satisfied with the new limit you can google on how to apply the throttlestop setting on each OS startup and then the values will always be applied.

Surprisingly enough, I have heard that the 2060, 2080, 2070 Super Max Q don't have as much of a temperature issue because the of the lower wattage of the 2060 and the 2070 max q card and the better cooling on the 2080. I have no idea if that is actually true though, it just seems that most people facing this issue have a fast Intel cpu and a 2070 super running at 115 watts.

Edit: Pictures!
Example of Throttlestop settings:
86423

Example of thermals over a 10 minute period I did as a quick demo with the new settings. These are lower than what I typically have because of the lower ambient temperature. GPU hovered around 67°C and CPU hovered around 77°C:
86424
Asus G732LWS
Geforce GTX 2070 Super
Intel i7-10875h

Desani wrote:
I would recommend attempting the adjustments with Throttlestop rather than Intel® Extreme Tuning Utility. Intel's utility is more user friendly but I found that it caused more system instability.

To replicate my settings in throttlestop you would need to do the following:
Run a quick benchmark with hardware info monitoring the CPU and GPU for an accurate before picture
Download and unzip throttlestop
Launch throttlestop and click the button Turn on at the bottom of the window. That activate throttlestop settings and turns off monitoring mode.
We are going to ignore most of the values as the BIOS does not allow you to modify some of the important values that other guides online tell you to change.
The only values that I changed are in TPL menu.
In the TPL menu the first two items are the long and short power max values. Change the long value from 45 to 35. Change the short value from 125 to 100. I left the turbo time limit alone. The value that is going to make the most difference is the long time value. I enabled clap on both of these settings, from what I have read clap tells the processor to attempt to keep the values below or at the value supplied.
Once the new values are inserted, click okay and with throttlestop turned on, the change is immediate.

Run another benchmark with the new settings and make sure it runs for at least a couple of minutes. You know notice that after the first spike that lasts 56 seconds, the temperature will slowly decrease over time and settle on a value that should be lower than before because of the wattage limit. On my laptop, that took the overall temperature from the high 80's and put it in the high 70's. It might vary from computer to computer so play with the values that keep the temperature to a value you are okay with. Once you are satisfied with the new limit you can google on how to apply the throttlestop setting on each OS startup and then the values will always be applied.

Surprisingly enough, I have heard that the 2060, 2080, 2070 Super Max Q don't have as much of a temperature issue because the of the lower wattage of the 2060 and the 2070 max q card and the better cooling on the 2080. I have no idea if that is actually true though, it just seems that most people facing this issue have a fast Intel cpu and a 2070 super running at 115 watts.


I'll have to try this tomorrow evening, but you are awesome man, I appreciate your research and guide. ASUS should look at your research and hopefully they could do something similar baked into the BIOS or something, IDK how that works.

Desani wrote:
I would recommend attempting the adjustments with Throttlestop rather than Intel® Extreme Tuning Utility. Intel's utility is more user friendly but I found that it caused more system instability.

To replicate my settings in throttlestop you would need to do the following:
Run a quick benchmark with hardware info monitoring the CPU and GPU for an accurate before picture
Download and unzip throttlestop
Launch throttlestop and click the button Turn on at the bottom of the window. That activate throttlestop settings and turns off monitoring mode.
We are going to ignore most of the values as the BIOS does not allow you to modify some of the important values that other guides online tell you to change.
The only values that I changed are in TPL menu.
In the TPL menu the first two items are the long and short power max values. Change the long value from 45 to 35. Change the short value from 125 to 100. I left the turbo time limit alone. The value that is going to make the most difference is the long time value. I enabled clap on both of these settings, from what I have read clap tells the processor to attempt to keep the values below or at the value supplied.
Once the new values are inserted, click okay and with throttlestop turned on, the change is immediate.

Run another benchmark with the new settings and make sure it runs for at least a couple of minutes. You know notice that after the first spike that lasts 56 seconds, the temperature will slowly decrease over time and settle on a value that should be lower than before because of the wattage limit. On my laptop, that took the overall temperature from the high 80's and put it in the high 70's. It might vary from computer to computer so play with the values that keep the temperature to a value you are okay with. Once you are satisfied with the new limit you can google on how to apply the throttlestop setting on each OS startup and then the values will always be applied.

Surprisingly enough, I have heard that the 2060, 2080, 2070 Super Max Q don't have as much of a temperature issue because the of the lower wattage of the 2060 and the 2070 max q card and the better cooling on the 2080. I have no idea if that is actually true though, it just seems that most people facing this issue have a fast Intel cpu and a 2070 super running at 115 watts.

Edit: Pictures!
Example of Throttlestop settings:
86423

Example of thermals over a 10 minute period I did as a quick demo with the new settings. These are lower than what I typically have because of the lower ambient temperature. GPU hovered around 67°C and CPU hovered around 77°C:
86424


I greatly appreciate your response and time. I will definitely try to do this myself. I hope that we can get any official information from ASUS support in this post.
Thanks a lot!

Desani wrote:
I would recommend attempting the adjustments with Throttlestop rather than Intel® Extreme Tuning Utility. Intel's utility is more user friendly but I found that it caused more system instability.

To replicate my settings in throttlestop you would need to do the following:
Run a quick benchmark with hardware info monitoring the CPU and GPU for an accurate before picture
Download and unzip throttlestop
Launch throttlestop and click the button Turn on at the bottom of the window. That activate throttlestop settings and turns off monitoring mode.
We are going to ignore most of the values as the BIOS does not allow you to modify some of the important values that other guides online tell you to change.
The only values that I changed are in TPL menu.
In the TPL menu the first two items are the long and short power max values. Change the long value from 45 to 35. Change the short value from 125 to 100. I left the turbo time limit alone. The value that is going to make the most difference is the long time value. I enabled clap on both of these settings, from what I have read clap tells the processor to attempt to keep the values below or at the value supplied.
Once the new values are inserted, click okay and with throttlestop turned on, the change is immediate.

Run another benchmark with the new settings and make sure it runs for at least a couple of minutes. You know notice that after the first spike that lasts 56 seconds, the temperature will slowly decrease over time and settle on a value that should be lower than before because of the wattage limit. On my laptop, that took the overall temperature from the high 80's and put it in the high 70's. It might vary from computer to computer so play with the values that keep the temperature to a value you are okay with. Once you are satisfied with the new limit you can google on how to apply the throttlestop setting on each OS startup and then the values will always be applied.

Surprisingly enough, I have heard that the 2060, 2080, 2070 Super Max Q don't have as much of a temperature issue because the of the lower wattage of the 2060 and the 2070 max q card and the better cooling on the 2080. I have no idea if that is actually true though, it just seems that most people facing this issue have a fast Intel cpu and a 2070 super running at 115 watts.

Edit: Pictures!
Example of Throttlestop settings:
86423

Example of thermals over a 10 minute period I did as a quick demo with the new settings. These are lower than what I typically have because of the lower ambient temperature. GPU hovered around 67°C and CPU hovered around 77°C:
86424



So one thing different is my default PP0 Current Limit is set to 200, and yours is 0, do I need to change that or does it not matter you think?

PC Benchmark only lowered my bench from around 7100 to 6900 after playing with these settings. Gonna try it in the real world a few days to see if we notice any difference.

Seems like this will be very helpful. Now ASUS really needs to look into this.

craig_cabbage wrote:
So one thing different is my default PP0 Current Limit is set to 200, and yours is 0, do I need to change that or does it not matter you think?

PC Benchmark only lowered my bench from around 7100 to 6900 after playing with these settings. Gonna try it in the real world a few days to see if we notice any difference.

Seems like this will be very helpful. Now ASUS really needs to look into this.


I looked into the pp0 value and I found this:

PP0 refers to the power limit of the CPU cores, and only the CPU cores. It does not include other things in the CPU such as cache, memory controller, integrated graphics and so on.

The PP0 Current Limit in ThrottleStop is set to 0. This usually means that this setting is not being used. Your CPU will only use as much current as it needs to use so setting this high is not going to hurt anything.


So it looks like on my laptop that setting is not using that setting but also having a setting of 200 might not affect too much as the CPU will only use as much as it needs. This might be a way to accomplish the same end goal of lowering thermals and could be explored further.
Asus G732LWS
Geforce GTX 2070 Super
Intel i7-10875h

Do you know how Armoury Crate and ThrottleStop interact? Should I set Armoury Crate to my desired power mode - silent/performance/turbo, and then turn on ThrottleStop, does it matter? Any insight?