cancel
Showing results for 
Search instead for 
Did you mean: 

ROG Strix G15 2021 G513QM - CPU Underperformance

NosliW
Level 8

Hi everyone,

I recently ran a 10-minute Cinebench R23 multicore stress test on my ASUS ROG Strix G15 (Ryzen 9 5900HX).  It looks like my CPU is underperforming. What are the reasons for this issue happening? Is this happening because of the pump-out effect of liquid metal?
Thanks in advance, and pardon me if I miss out anything.

Windows OS: Windows 11
Armoury Crate is set to performance mode
Screenshot 2025-06-02 121112.pngScreenshot 2025-06-02 121159.pngScreenshot 2025-06-02 125258.png

 

254 Views
1 ACCEPTED SOLUTION

Accepted Solutions

ElectroStingz
Level 13

That's a shame as it would have been good to know the before scores, CPUs can degrade over time, software and drivers also have an impact / OS updates so if the test environment is completely different one must factor this but one would expect somewhere close to 10k at least. 

Turbo mode would also cause the GPU to run at a higher boost, your RTX 3060 max temp spec is 93 Deg C so you are over that which will cause the GPU to try and cut back (thermal throttling which causes stutters). The GPU also has another temperature, GPU Hotspot temperature so if your GPU core is 94, the hotspot is probably peaking over 105+, which is not good for it.

Try hardware monitor for more info / sensors readings if you haven't used already.

https://www.cpuid.com/downloads/hwmonitor/hwmonitor_1.57.zip

As to the CPU core temperature differences what did you mean by the tilt having an impact?

Your 5900HX has 2 Core Dies that make direct contact with the heatsink, so if a certain group of cores (odd vs even) show a higher temp reading it always means the paste has shifted.

The 2 dies at the top are your Cores.

5900xTIM_1.jpg

Orange is thermal paste when new

5900xTIM_2.jpg

This is thermal paste over time, if it's like this now where one core is significantly less thermal paste coverage it explains your 10-15 deg C difference on odds vs evens.

5900xTIM_3.jpg

But given your current testing and readings plus the age of the laptop it is a good idea to get things serviced and all the paste replaced. CPU, GPU, VRM (all the mosfets and power inductors (thermal putty required)).

View solution in original post

4 REPLIES 4

ElectroStingz
Level 13

Hello,

Do you have any previous runs on the same laptop that you are using for comparison that were done when the laptop was new? 

Power consumption looks a little low and your core temperatures look slightly uneven

Cores 0,2,4,6 are around 10-15 Degrees C hotter than 1,3,5,7. This would usually suggest a hotspot on the die / issue with thermal material.

And your Vcore seems quite low.

Are you running a modified configuration?

Hi, thanks for your reply!

Unfortunately, I didn’t run any benchmarks (Cinebench, FurMark, etc.) when I first received the laptop, as this is my first gaming laptop and I was completely new to this kind of testing at the time.

Regarding the uneven core temperatures, it might partly be due to the fact that I usually place the laptop on a tilted stand, which could potentially affect heat dissipation. But now I keep it flat and use a cooling pad.

I haven’t made any manual tweaks to voltage or clock settings. Armoury Crate is set to Performance mode, and everything else remains at stock. There’s no BIOS undervolting, no custom fan curve, and no third-party tuning software involved. The only change is that AMD Radeon Adrenalin has the automatic tuning option set to “CPU Overclocked.”

I also want to mention something else I’ve observed:
When I run FurMark in Performance mode, everything seems to work fine. But when I switch to Turbo mode, the benchmark becomes noticeably laggy, with stuttering every 2–3 seconds.

Could this be potentially due to a bad power adapter?

Performance modePerformance mode

 

 

Turbo modeTurbo mode

ElectroStingz
Level 13

That's a shame as it would have been good to know the before scores, CPUs can degrade over time, software and drivers also have an impact / OS updates so if the test environment is completely different one must factor this but one would expect somewhere close to 10k at least. 

Turbo mode would also cause the GPU to run at a higher boost, your RTX 3060 max temp spec is 93 Deg C so you are over that which will cause the GPU to try and cut back (thermal throttling which causes stutters). The GPU also has another temperature, GPU Hotspot temperature so if your GPU core is 94, the hotspot is probably peaking over 105+, which is not good for it.

Try hardware monitor for more info / sensors readings if you haven't used already.

https://www.cpuid.com/downloads/hwmonitor/hwmonitor_1.57.zip

As to the CPU core temperature differences what did you mean by the tilt having an impact?

Your 5900HX has 2 Core Dies that make direct contact with the heatsink, so if a certain group of cores (odd vs even) show a higher temp reading it always means the paste has shifted.

The 2 dies at the top are your Cores.

5900xTIM_1.jpg

Orange is thermal paste when new

5900xTIM_2.jpg

This is thermal paste over time, if it's like this now where one core is significantly less thermal paste coverage it explains your 10-15 deg C difference on odds vs evens.

5900xTIM_3.jpg

But given your current testing and readings plus the age of the laptop it is a good idea to get things serviced and all the paste replaced. CPU, GPU, VRM (all the mosfets and power inductors (thermal putty required)).

Hi, thanks a lot for the detailed explanation.

I’ll go ahead and have the system serviced and repasted — including the CPU, GPU, and VRMs as you suggested. It definitely sounds like it’s due for maintenance at this point.

And yes, what I meant by the tilt was exactly as you described. My laptop has been used on a stand at a tilted angle for an long time, so I was thinking that could have caused the liquid metal to gradually shift over time.

Thanks again for the valuable insight!