1.3.5

Improvements

• Only GPUs connected to a display device will be considered for optimizations by default
  • this behavior is controlled by the new optimizer service configuration property gpu.only_connected
  • this behavior is only implemented for AMD GPUs at the moment

1.3.4

Improvements

• do not try to disable window compositors when compositor.off if requested from a wayland session
• do not try to hide the mouse pointer when mouse.hidden is requested from a wayland session (unclutter does not support wayland at the moment)
• adding a new project definition/setup (pyproject.toml) file to comply with the latest standards

1.3.3

Improvements

• optimizer: ignoring more unneeded EA launcher and Wine child processes

Fixes

• watcher/optimizer: not able to map processes alive with ps command version >= 4.0.X

1.3.2

Improvements

• replaced some subprocess calls executed in behalf of non-root users by async calls (Python's asyncio native approach)

Fixes

• the optimizer service (as root) not able to execute some commands in behalf of non-root users (started with Python 3.10.9)

1.3.1

Improvements

• Optimizing the children of the target process
• New optimizer service properties:
  • optimize_children.timeout: maximum period in seconds to keep looking for the target process children (default: 30). 0 can be defined if children should be ignored.
  • optimize_children.found_timeout: maximum period in seconds to still keep looking for the target process children after a child in found (default: 10). 0 can be defined if the search process should be interrupted immediately after a child is found.
• Ignoring some Ubisoft launcher sub-processes not required to be optimized (when launched from Steam)

Fixes

• Only checking for mapped processes when a process optimization is requested

1.3.0

Improvements

• Properly detecting all Steam games subprocesses that need to be optimized (no need to map launchers anymore)
• Launcher mapping now optimizes all matches instead of just the last born process
• New optimizer service property launcher.mapping.found_timeout: maximum time in seconds to still keep looking for a process mapped to a different process after a match. This property also affects the period to look for Steam subprocesses. (default: 10)
• The optimizer service property launcher.mapping.timeout has now a default value of 60 (seconds)

Fix

• wild card mapping to proper regex pattern

1.2.2

Improvements

• Minor code refactoring and log improvements regarding AMD GPU management
• Optimizer:
  • configuration property check.finished.interval now accepts floats and the minimum value accepted is 0.5
  • new configuration property gpu.id: allows to define which GPU cards should be optimized (e.g: gpus.id = 0,1). If not defined, all available GPUs are considered (default).

Fixes

• optimizer:
  • when running as a system service, sometimes the GPU mapped directories are not available during the system startup and affects the correct behavior of the property gpu.cache when it is enabled (true)
    • so now the available GPUs will be cached after a first request when the optimizer is running as a system service (otherwise they will be cached normally during the service startup process)

1.2.1

Fixes

• Performance mode not being activated for AMD GPUs from the RX 6XX0 series (tested on kernels >= 5.15)

Improvements

• removed unused code

1.2.0

Features

• watcher service:
  • allowing processes to be ignored through the mapping file: watch.ignore (must be located in ~/.config/guapow or /etc/guapow)
    • it follows the same patterns as the watch.map file, but the profile name is not required (as it makes no sense). e.g:

    my_app_name
    my_app_name*  
    /bin/my_proc
    r:/bin/.+/xpto
    

    • this feature is useful if you have general mappings that cover a lot of processes in watch.map (e.g: /usr/bin/*), but want to ignore specific ones
    • new config property ignored.cache to cache all mapped patterns to memory after the first read and skip next I/O calls (default: false)

1.1.1

Fixes

• AMD GPU performance mode not working #1
  • (test context -> GPU: Ryzen 7 5700G, Kernel: 5.15.45, O.S: Arch Linux)