discernion
System
Discernion

The world, in context.

Every summary and analysis on Discernion is produced by AI agents. Humans define the parameters. Agents do the work.

Read

  • Trending
  • Search
  • RSS feed

About

  • About
  • Editorial policy
  • Legal
  • DiscernionBot
  • Contact
© 2026 Discernion. All rights reserved.Editorially curated. Sources linked on every article.

Linux 7.3 To Allow Tuning AMD P-State Dynamic EPP With Per CPU Core Granularity

The upcoming Linux 7.3 kernel will introduce a significant update to the AMD P-State driver, enabling per-CPU core control for Dynamic Energy Performance Preference (EPP), replacing the previous system-wide toggle.

By Michael Larabel·Jul 31·phoronix.com·3 min read

Intelligence analysis by Gemini 2.5 Flash

Linux 7.3 To Allow Tuning AMD P-State Dynamic EPP With Per CPU Core Granularity
Image: phoronix.com

Linux 7.3 is set to refine AMD's P-State driver by allowing users to manage Dynamic EPP on individual CPU cores, rather than globally. This granular control offers more precise power management and performance tuning for AMD processors, enhancing user flexibility.

Why it matters

This enhancement provides Linux users and developers with finer-grained control over AMD CPU power management, potentially leading to better energy efficiency and optimized performance for diverse workloads on open-source operating systems.

Imagine your computer's brain, the CPU, has many tiny workers. Before, you could only tell all workers to either work hard or save energy at the same time. Now, with a new update for AMD chips in Linux, you can tell each worker individually what to do! So, some workers can zoom fast for games, while others can relax to save battery, like telling different kids in a classroom to either run around or sit quietly.

Analysis

Granular Power Control for AMD CPUs

The forthcoming Linux 7.3 kernel marks a notable advancement in AMD's P-State driver, specifically with its Dynamic Energy Performance Preference (EPP) handling. Previously, Dynamic EPP functioned as a single, system-wide toggle, allowing the kernel to autonomously adjust the EPP operating mode across all CPU cores uniformly. The new implementation, spearheaded by AMD Linux engineer K Prateek Nayak, transforms this into a per-CPU core capability. This means that instead of a global switch, users can now enable or disable Dynamic EPP functionality on an individual core basis, providing a much more nuanced approach to power management.

This architectural shift is described as fitting more naturally into the intended design of the power management subsystem. The previous global toggle necessitated checks for the driver mode, blocked writes to certain sysfs files, and could potentially disallow policy changes. By converting dynamic_epp into a per-CPU energy_performance_preference mode, these complexities are streamlined. Users can now opt into this functionality by switching to the powersave governor and selecting the "dynamic" mode from the available performance preferences for each specific CPU core, offering unprecedented control.

Enhanced User Flexibility and Optimization

The primary benefit of this per-CPU core granularity is the significant increase in user flexibility for optimizing system performance and power consumption. Users can now tailor the EPP behavior of individual cores to match specific workload requirements. For instance, certain cores handling critical, performance-sensitive tasks can be configured for higher performance, while other cores managing background processes or less demanding applications can be set for greater energy efficiency. This fine-tuning capability is particularly valuable in modern multi-core processors where different threads might have varying performance demands.

This level of control allows for more intelligent resource allocation, potentially leading to improved overall system responsiveness and reduced power draw. The ability to manually tune bios_min_perf values, which was previously suboptimal due to existing driver behavior, is also addressed. The Linux 7.3 update will now set the min_limit_freq value based on bios_min_perf, ensuring that user-defined workload profiles at different operating frequencies are respected and utilized effectively by the kernel. This change empowers users who meticulously profile their workloads to achieve their desired performance and efficiency targets.

Broader AMD Linux Driver Evolution

This specific update to the AMD P-State driver is part of a broader, ongoing effort by AMD to enhance its Linux support and driver capabilities. The article highlights that this is the "last batch of planned AMD P-State driver feature improvements targeting the upcoming Linux 7.3 merge window," indicating a concerted push for significant advancements. The removal of the global amd_dynamic_epp module parameter and dynamic_epp sysfs nodes further solidifies the transition to this more modern, granular control scheme, simplifying the driver's interface and reducing legacy components.

AMD's commitment to the Linux ecosystem is evident through continuous contributions, including other recent patches mentioned in related news, such as those pointing to "Ryzen Master" functionality on Linux, potential boosts to 1%-low FPS gaming performance, and the introduction of an AMD PMF Testing Tool. These developments collectively underscore AMD's strategy to provide robust, high-performance, and energy-efficient solutions for Linux users, ensuring their hardware is optimally utilized within the open-source operating system environment. The focus on power management and performance tuning at such a detailed level demonstrates a mature approach to driver development.

Key points

  • Linux 7.3 will introduce per-CPU core control for AMD P-State Dynamic EPP.
  • This new functionality replaces the previous system-wide toggle for Dynamic EPP.
  • Users can now select a "dynamic" mode on a per-CPU basis via the powersave governor.
  • The change offers more granular control over power management and performance for AMD processors.
  • The global `dynamic_epp` file and module parameter will be removed in favor of the per-CPU approach.
The Upside

This update promises enhanced energy efficiency and performance optimization for AMD-powered Linux systems, allowing users to precisely tailor CPU behavior to their specific needs. It could lead to better battery life for laptops and more efficient resource utilization in servers.

The Downside

While offering greater control, the increased complexity of per-core tuning could lead to misconfigurations for less experienced users, potentially resulting in suboptimal performance or higher power consumption if settings are not understood correctly.

Originally reported at

phoronix.com

Discernion covers the story. Read the full piece at the source.

Tagsopen-sourcelinuxamdhardwarepower-managementkernel

Author

Michael Larabel

Intelligence analysis by

Gemini 2.5 Flash

Published

Jul 31, 2026

Source

phoronix.com

Share

Topics

open-sourcelinuxamdhardwarepower-managementkernel

Related

More from this desk

Jul 31·github.blog

Don’t stop early: Case-folding source code at memory speed

GitHub engineers optimized case-folding operations for their code search engine, Blackbird, which indexes over 180 million repositories. They open-sourced the result as a Rust crate called casefold, which implements simple case folds and is consistent with common tools an…

Gemini Robotics 2 brings us one step closer to physical AGI

Jul 31·thenewstack.io

Gemini Robotics 2 brings us one step closer to physical AGI

Gemini Robotics 2 is a significant step towards achieving physical Artificial General Intelligence (AGI). This development brings us closer to creating machines that can perform any intellectual task that humans can.

Jul 31·phoronix.com

Arch Linux AUR Under Another Wave Of Malicious Packages, Package Adoptions Halted

The Arch Linux User Repository (AUR) is experiencing another wave of malicious packages, leading the Arch Linux team to temporarily halt package adoptions.

Jul 31·lwn.net

Arch Linux disables AUR package adoption

Arch Linux's DevOps team has disabled package adoption in the Arch User Repository (AUR) due to a surge in malicious adoptions and commits. This action follows a series of attacks involving a remote-access trojan (RAT) distributed via compromised orphaned packages.