Global CTD-SLT processing occurs at level four of the transputer trigger network. One extra T800 TRAM card is dedicated to this SLT level and may reside on an additional ROC board in the regional first level trigger box.
The processor's main function is to process the track information from the 16 track finding processors (at level three) to form the global CTD-SLT summary data. This summary data comprises event information and parameters for the found tracks. The online bank format can be found in Reference 7 and the offline format in the ADAMO subschema GSLCTD of the ZEUS DDL.
Before global processing can begin a few network tasks must be performed. The global SLT transputer closes the chain of track finding processors. As such, it must pass vector hits from crate 0 around to crate 15 to allow each track finding process to form two-crate input masks for pattern recognition. To allow identical track finding code to run on all processors at level three, an empty packet of track data must be passed to crate 0. Since concatenation of track data occurs on each node in level three, crate 15 will pass tracks from the entire CTD to level four.
Some of the tracks passed along the chain may share vector hits because of the overlapping input masks. These duplicate tracks are rejected before global processing begins. At this point track multiplicity and z-vertex cuts may be applied to reduce the data size sent to the GSLTB and processing time at level four.
After the global processing is finished the data is transfered to the GSLTB using a dedicated link with the protocol described in Reference 7 running on it. In addition, the data is sent to the global CTD readout transputer via the harness link. The readout transputer adds the ADAMO header and passes the data to the EVB every GSLT accept. The begin-of-run and end-of-run data is readout in a similar way to the normal trigger summary data.
Besides this regular processing, level four is responsible for communicating acknowledge signals back to run control via the harness. The global SLT processor acknowledges the setup and begin-of-run states.