investigate event driven vs time-based looping #25
Labels
enhancement
New feature or request
Comments
swiftraccoon
pushed a commit
that referenced
this issue
Dec 1, 2023
Upgraded the project's C++ standard requirement to C++20 for improved language features. Introduced Boost libraries, specifically for system and thread components, to enhance asynchronous processing capabilities and added the inotify-cpp submodule for efficient file system monitoring. These changes aim to refine file handling operations such as file processing and transcription by leveraging modern C++ paradigms and libraries. Additionally, the ConfigSingleton has been updated to include new thread management options and resolve naming inconsistencies. This improvement potentially increases the application's performance, particularly in handling file events and concurrent transcriptions. In the update to `fasterWhisper.py`, code has been organized with better structure, added argument parsing, and improved logging. It also includes the model's initialization method and error handling for audio transcription, reflecting a significant refactoring for better maintainability and readability. Modified the related configurations and build scripts to integrate these changes and guarantee seamless building with the new dependencies. Ref: integrate parallel processing #23 , investigate event driven vs time-based looping #25
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
presently we accept a millisecond configuration value to loop
main.cpp/processDirectory-- is it significantly better to do file-based events to triggerprocessDirectory?even at 100ms loop there is basically no CPU load tho (~0%).
if we do remove the milliseconds and just do
while (true)we get to ~6% CPU usageThe text was updated successfully, but these errors were encountered: