Thinking ahead to memory fragmentation and tools running for days at a time...
m_time_int is a vector used in NHits & TestVertices to take a SubSample's (nominally double) times and convert them to an int. In order to prevent memory fragmentation (and for speed), it's a good idea to allocate as much memory as required in the Initalise() method, so that the vector isn't moved around in memory multiple times as it sees larger and larger SubSamples
Studies should be done on an appropriate value - but having the SubSample width used by PrepareSubSamples available in the data model would be useful (so you can potentially do something like 1.5 * average_dark_noise_rate_in_subsample_width)
Thinking ahead to memory fragmentation and tools running for days at a time...
m_time_int
is a vector used in NHits & TestVertices to take a SubSample's (nominally double) times and convert them to an int. In order to prevent memory fragmentation (and for speed), it's a good idea to allocate as much memory as required in theInitalise()
method, so that the vector isn't moved around in memory multiple times as it sees larger and larger SubSamplesStudies should be done on an appropriate value - but having the SubSample width used by PrepareSubSamples available in the data model would be useful (so you can potentially do something like
1.5 * average_dark_noise_rate_in_subsample_width
)