I cannot be absolutely sure if it’s a regression only in NVIDIA drivers, and not due to the fact that I’ve also upgraded Linux 5.11 to 5.12 but here’s what I have now: Idle power consumption NVIDIA drivers 465.27/Linux kernel 5.12.X: 10.3W NVIDIA drivers 465.24/Linux kernel 5.11.X: 8.5W Everythi…

Power consumption regression in NVIDIA Linux drivers 465.27 [cannot reproduce]

birdie May 20, 2021, 4:26pm 2

I cannot reproduce it currently. Let’s mark it INVALID for a while.

Topic		Replies	Views
A slight idle power consumption increase/regression recently Linux	2	1318	February 13, 2022
Increased idle consumption with driver 570 Linux	25	3349	June 23, 2025
Very(!) slow ramp down from high to low clock speeds leading to a significantly increased power cons Linux	159	26414	February 6, 2024
Insufficient power gating for high resolution/high refresh rate modes in Linux vs. Windows Linux	4	709	September 25, 2022
Linux drivers since 470 running at higher idle power consumption Linux power	0	367	March 27, 2023
Extremely power consumption after nvidia-suspend with 4060 laptop Linux	1	183	October 27, 2024
Nvidia smi driver version 570 on ubuntu 24.04 shows wrong power consumption when running for the first time Drivers - Linux, Windows, MacOS	0	192	March 8, 2025
Petition: extend support for 384.XX drivers (given an idle power consumption regression in new drivers) Linux	1	563	July 18, 2018
If you have GPU clock boost problems, please try __GL_ExperimentalPerfStrategy=1 Linux	57	27998	June 27, 2023
NVIDIA driver has a habit of keeping my GPU at the highest performance level Linux	24	6521	February 17, 2025