Unprecedented data rates that are expected at the LHC put high demand on the speed of the detector data acquisition system. The CSC subdetector located in the Muon Endcaps of the CMS detector has a data readout system equivalent in size to that of a whole Tevatron detector (60 VME crates in the CSC DAQ equal to the whole D0 DAQ size). As a part of the High Level Trigger, the CSC data unpacking runs online and it needs to be able to cope with high data rates online. Early versions of the unpacking code used bit shifts and masks to unpack binary data. To reduce the unpacking time we decided to switch to bit field based data unpacking. The switch allowed us to gain an order of magnitude in speed. In this paper we explain how bit field data unpacking works and why it is dramatically faster compared to conventional bit shift and mask methods.
展开▼
