Events in the partitions 230, 240, 250 timed out during reconstruction. Partition 7796.260 took a long time but finally finished. Looking at the output it became clear that the SVT occupancy was very high. For instance, here with over 5000 SVT Raw Tracker Hits

Leading to events such as the following:


Here is a plot of the normal number of SVT raw hits and a plot of reconstruction time versus size:


Evidently something has gone wrong with the SVT operating conditions. Propose adding org.hps.recon.filtering.EventFlagFilter to the steering file. This checks on a number of SVT-related quality flags and aborts further processing if any of them are set.

However, this appears to also skip 40% of what are most likely good events. They are failing the svt_latency_good flag. This needs to be checked.


  • No labels