rtac.utils.background_thread_control

Numpy, scipy and sklearn background libraries spawn threads that may clog CPU through oversubscription. This can interfere with the performance measurement of algorithms that are configured, especially if all system cores are utilized in the tournaments. Because of this, we set the maximum thread numbers to 1. It has to happen before any of the modules are loaded.

Functions

set_background_thread_nr()

Set thread number of background libraries for numpy, scipy, sklearn to 1.

rtac.utils.background_thread_control.set_background_thread_nr() None[source]

Set thread number of background libraries for numpy, scipy, sklearn to 1.

Return type:

None