The Design And Research On The Framework Of Event Selection Software Of BEijing Spectrometer Ⅲ

The BESIII is a general purpose experiment at BEPCII collider, which is currently under construction in Beijing. The main character of BEPCII is multiple bunch, high luminosity and the extremely high data rate. This thesis concerns the BESIII software R&D environment and the BESIII online Event Filter, which is developed in the BESIII software R&D environment.The BESIII software R&D environment uses Linux system and object oriented C++ language. The underlying software framework is Gaudi, which is also used by many other HEP experiments. The Configuration Management Tool (CMT) and Concurrent Versions System (CVS) are adopted for the configuration management of our software. All of these make the whole software R&D process under control, programs more reliable and also ensure the software development in an efficient way.The BEPCII is a multiple-bunch machine. When it works on the designed luminosity, the event rate is 107Hz. The trigger system is divided into two levels so that we can record the physics event that we are interested in. The Level 1 is a hardware trigger, which uses front-end electronic (FEE) to do the real time event selection, in order to suppress the event rate to about 4 kHz. The Level 2 is an online event filter, which uses the computer farm to do the further suppression, to reduce event rate to 3 kHz. We studied many other HEP experiment's trigger systems, decided to use one software framework for both online event filter and offline data processing. The online event filter software is designed to have two parts: 1) the event selection, which can run in the offline data processing system independently. 2) The event selection controller, which is the interface between event selection and processing task. By using this method, we can reuse the offline software for online event selection, so that we can reduce the cost and improve the system's maintainability. From the test results, we find the online event filter software, which is developed based on Gaudi framework, can work properly. This will be a good foundation for the further test and performance tuning.