v0.4.4

[0.4.4] - 2023-09-12

Bug Fixes

• Avoid setting power cap if it was not changed
• Inability to add points to an empty fan curve (#189)

Features

• Standardise cli arguments, configure gui log level through cli args
• Show vulkan driver info in gui
• Print vbios date from drm

Miscellaneous Tasks

• Clippy fixes
• Update README screenshots
• Bump dependencies
• Update dependencies, adopt breaking changes

Packaging

• Add prebuilt arch package (#192)

v0.4.3

Changelog

All notable changes to this project will be documented in this file.

[0.4.3] - 2023-05-14

Bug Fixes

• Check for DEVTYPE=drm_minor when trying to find usable gpu paths instead of using the name
• Bump package version

Features

• Retry GPU initialization when no GPUs were found
• Settings confirmation (#163)
• Report gui version
• Parallel ci (#178)

Miscellaneous Tasks

• Bump dependencies
• Switch to stable vulkano release
• Bump dependencies and switch to a tagged libdrm_amdgpu_sysfs version

v0.4.2

Features

• Add support for not cleaning up clocks table on shutdown (#133)
• Avoid setting clocks value unless it was changed

v0.4.1

Features

• Add DRM info (#132)

Bug Fixes

• Don't show minimum value scales when data is not available

Miscellaneous Tasks

• Bump version

Packaging

• Add Fedora 38 package

v0.4.0

Features

• Reconnection to daemon socket if it closes
• Add support for configuring the minimum clockspeed and voltage

Miscellaneous Tasks

• Bump dependencies

v0.3.2

Features

• Reload configuration on dbus suspend/resume events

Miscellaneous Tasks

• Update README

v0.3.1

Features

• Support for power profile modes
• Daemon implementation for creating module conf file with ppfeaturemask set
• Gui for enabling overclocking

Documentation

• Update README to include overclocking info

Miscellaneous Tasks

• Bump version
• Bump tokio from 1.25.0 to 1.26.0 (#111)

Refactor

• Reexport amdgpu-sysfs from lact-schema to avoid having to specify the version twice

v0.3.0

This release includes several major changes:

• Voltage offset support and fixes for RX 6000 series GPUs

  Due to format differences in the GPUs of this series, it was not possible to properly configure clocks settings on them. After debugging and help from various people in #51 this has been resolved and now it's additionally possible to configure the voltage offset value (not present on older GPUs)

• UI redesign

  The UI has been updated to look cleaner and have consistent padding around elements, see #106 for side-to-side comparison. Thanks @treuks for helping with the design.

Minor changes:

• The daemon no longer fails to start if it can't apply the current settings, now it prints an error instead. This can happen for example if you've disabled overdrive support after having it previously enabled.
• There are more logs on the "debug" logging level (disabled by default)

v0.2.1

This is a hotfix release that fixes a deadlock when switching between GPUs on a multi-GPU system.

v0.2.0

This release has a LOT of changes, including:

• The GUI now uses Gtk4
• The daemon has been rewritten and now exposes a JSON API over a conventional unix socket, see here for more
• All of the SysFS handling is now done as part of the amdgpu-sysfs crate, which now fully supports both new and old formats for the clocks table, leading into the next point
• You can now change voltage on GPUs using the new clocks table format (Vega20+)
• The fan curve can now be fully customized from the UI, with the ability to add/remove points and edit temperatures
• There are now searchable lists of Vulkan features and extensions
• LACT now has an icon! It was created by @NalianNalis, see #15
• There is now a workflow to build .deb and .rpm packages using pkger

Unfortunately, due to config format changes (and GPU identification in particular) your settings will not be saved when updating from an older version.

There are a lot of internal changes as well. This is a large release which addresses a lot of the issues that were present before, and will make it possible to implement more features that were not present before.

Please report any issues that you get, as it is not possible to test every existing hardware configuration.