The Simplest Way to Control Nvidia GPU Fan Speed in Linux
The GitHub repository provides a script for controlling NVIDIA GPU fan speed in Linux. It includes a fan curve graph and instructions for setup, dependencies installation, and script execution to manage fan speed efficiently.
Read original article
The GitHub repository offers a script designed for managing NVIDIA GPU fan speed in Linux. It features a fan curve points graph and detailed guidance on configuring the environment, installing necessary dependencies, and executing the script. By following the instructions provided in the repository, users can effectively regulate the fan speed of their NVIDIA GPU on a Linux system using the script.
Related
Show HN: Pdfscale
The GitHub repository hosts "pdfScale," a Bash script using ghostscript for PDF scaling and resizing via the command line. It supports various modes, paper sizes, and installation methods. Find more details on the repository.
Neko: Portable framework for high-order spectral element flow simulations
A portable framework named "Neko" for high-order spectral element flow simulations in modern Fortran. Object-oriented, supports various hardware, with detailed documentation, cloning guidelines, publications, and development acknowledgments. Additional support available for inquiries.
Ubuntu 24.10 Now Defaults to Wayland on Nvidia
Ubuntu 24.10 will switch to using Wayland over X11 for NVIDIA GPUs due to improved driver support. This aligns with GNOME 47 updates, enhancing performance and user experience with modern graphics technologies.
Guest Users Module for Linux
The "Guest Users Module for Linux" on GitHub supports guest user creation on Linux using PAM and GNU nss. It ensures user separation, offers installation guidance, configuration options, and development support.
Show HN: Code to run Gemini (Nano) locally on desktop/Chrome
The GitHub guide explains running Google Nano on desktop with Chrome Canary. It includes setup steps, testing guidance, and a demo link for practical exploration. Find detailed instructions on the GitHub page.
4 commentsStupidest? Definitely.
What one really wants is a target temperature. There are two pieces I'd use here:
- Baseline open-loop controller based on power usage (integrated over some time; e.g. past 30 seconds)
- Closed-loop controller for fine-tuning with feedback from measured temperature to fan speed (PID is fine, but the integral term is critical)
The latter can be used to periodically calibrate the former.
You do not want a simple mapping of temperature to fan speed.
The odd issue about the above is that done wrong, continuously-changing fan speeds can be upsetting to users. Once can use a discrete approximation with hysteresis to avoid that.
Related
Show HN: Pdfscale
The GitHub repository hosts "pdfScale," a Bash script using ghostscript for PDF scaling and resizing via the command line. It supports various modes, paper sizes, and installation methods. Find more details on the repository.
Neko: Portable framework for high-order spectral element flow simulations
A portable framework named "Neko" for high-order spectral element flow simulations in modern Fortran. Object-oriented, supports various hardware, with detailed documentation, cloning guidelines, publications, and development acknowledgments. Additional support available for inquiries.
Ubuntu 24.10 Now Defaults to Wayland on Nvidia
Ubuntu 24.10 will switch to using Wayland over X11 for NVIDIA GPUs due to improved driver support. This aligns with GNOME 47 updates, enhancing performance and user experience with modern graphics technologies.
Guest Users Module for Linux
The "Guest Users Module for Linux" on GitHub supports guest user creation on Linux using PAM and GNU nss. It ensures user separation, offers installation guidance, configuration options, and development support.
Show HN: Code to run Gemini (Nano) locally on desktop/Chrome
The GitHub guide explains running Google Nano on desktop with Chrome Canary. It includes setup steps, testing guidance, and a demo link for practical exploration. Find detailed instructions on the GitHub page.