Unsolved
1 Rookie
•
3 Posts
0
69
March 3rd, 2026 12:13
GPU Power Limit Restricted to 50W in Ubuntu While Windows Allows 95W – Firmware Clarification Requested
I have an Alienware 16X Aurora AC16251 with an RTX 5070 Laptop GPU.
Under Windows, when using Performance Mode in Alienware Command Center, the GPU power consumption reaches approximately 95W under load.
However, when booting into Ubuntu 24.04 (dual boot setup), the GPU power limit appears capped at 50W under identical load conditions.
In Ubuntu, nvidia-smi reports:
Maximum power limit: 115W
Default/current power limit: 50W
Attempts to change the limit are not supported
This suggests the GPU is capable of a higher power level, but it is being restricted outside of Windows.
I would like to understand:
Is this behavior expected by design?
Is the higher GPU TGP dependent on Windows software (Alienware Command Center)?
Is there a BIOS or firmware setting that enables the full advertised GPU TGP independent of operating system?
I am not requesting Linux support specifically — I am trying to understand whether the GPU power configuration is OS-dependent by design.


ejn63
10 Elder
•
30.8K Posts
1
March 3rd, 2026 14:30
It would be a function of the software (driver), not the hardware or the firmware. Since Dell does not support Linux at all on Alienware notebooks, it would have no reason to limit the power to the GPU -- it's up to the Linux community to come up with a driver that supports what you want to do.
Tesla1856
8 Wizard
•
17.5K Posts
0
March 3rd, 2026 17:44
In Ubuntu, you might try installing the (optional) proprietary Nvidia drivers (if they also support mobile-GPU's).
Your Ubuntu is a bit old. My experience is with v25 on desktop-class hardware.
(edited)
Tarpit
1 Rookie
•
3 Posts
0
March 5th, 2026 07:30
@ejn63
Thank you for the response.
I understand that Dell does not officially support Linux on Alienware systems, and that driver support on Linux is primarily handled by NVIDIA and the Linux community.
However, what led me to question whether this might be platform-level behavior is that nvidia-smi in Ubuntu reports a maximum power limit of 115W, while the default/current limit remains fixed at 50W, and attempts to change it with nvidia-smi -pl are rejected as “not supported”.
This suggests that the driver is able to detect the GPU’s higher power capability, but the power limit itself cannot be modified from the operating system. Under Windows, when using Alienware Command Center in Performance Mode, the GPU reaches around 95W under load, which indicates that the platform is capable of running the GPU at a higher TGP.
Because of that difference, I was trying to determine whether the higher power profile might be activated through system power profiles (for example through Alienware Command Center) rather than directly through the NVIDIA driver.
My main goal with this post is simply to understand whether the system’s GPU power configuration is tied to those Windows power profiles or if there is any BIOS/firmware configuration that controls it independently of the operating system.
Tarpit
1 Rookie
•
3 Posts
1
March 5th, 2026 07:49
@Tesla1856
Thank you for the suggestion.
I did try both the open and proprietary NVIDIA drivers in Ubuntu. Initially the system was using nvidia-driver-590-open, and I also tested the proprietary nvidia-driver-590. In both cases the behavior remained the same: the GPU power limit stays at 50W in Ubuntu.
When checking with nvidia-smi, the tool reports a maximum power limit of 115W, but the default/current limit is fixed at 50W, and attempts to change it with nvidia-smi -pl return that the operation is not supported. This is what made me think the higher power profile might be controlled at the system/platform level rather than directly through the driver.
Regarding the Ubuntu version, I am currently using Ubuntu 24.04 LTS, which is the current long-term support release. Since the GPU is detected correctly and the driver loads normally, I suspect the power limit behavior may not be related to the Ubuntu version itself.
My main goal with this post is to understand whether the higher GPU power profile that Windows enables (around ~95W under load using Alienware Command Center) depends on those Windows power profiles, or if there is any system-level configuration that allows the same behavior independently of the operating system.