cvpy.utils.CASThreadTuner.CASThreadTuner.tune_thread_count¶

static CASThreadTuner.tune_thread_count(action_function: Callable[[CAS, ndarray, ndarray], float], setup_function: Callable[[], CAS], teardown_function: Callable[[CAS], None], iterations: int = 5, controller_thread_range: range = range(4, 65, 4), worker_thread_range: range = range(4, 65, 4), objective_measure: Statistic = Statistic.MEAN) → CASThreadTunerResults¶

Compute the optimal thread count for a given image action.

Parameters:

action_function:class:’function’: Specifies a user-defined function that calls an image action.
setup_function:class:’function’: Specifies a user defined function to set up CAS environment
teardown_function:class:’function’: Specifies a user defined function to terminate the CAS session.
iterations:class:’int’: Specifies the number of iterations to call action_function for each combination of threads.
controller_thread_range:class:’range’: Specifies the range of threads on controller node.
worker_thread_range:class:’range’: Specifies the range of threads on each worker node.
objective_measure:class:’enum.EnumMeta’: Specifies the objective measure for performance over the given iterations - mean, median, minimum, maximum, stdev.

Returns:

class:: ‘__main__.CASThreadTunerResults’ ..