Speaker
Description
Summary
The Global Calorimeter Trigger (GCT) is the last stage of the
calorimeter trigger
chain. The primary purpose of the GCT is to reduce the
number of calorimeter
trigger objects that need to be considered by the Global
Trigger (GT) for a First
Level Trigger Accept (L1A) decision. The pipeline memories
that store event
information prior to a L1A request have only limited depth
and thus data
transmission and processing time must be kept to a short,
critical time period.
The trigger objects sent to the GT are listed below.
The "rank" of an electron or
jet is at present its transverse energy, however in principle
it could also be
derived from jet location and energy. The jet transverse
energy is the sum of both
the hadronic and electromagnetic calorimeter.
- 4 isolated and 4 non-isolated electrons of highest rank
- 4 central, 4 forward and 4 tau clustered jets of highest
rank - total transverse energy: sum of all jet transverse energy
(magnitude) - missing transverse energy (magnitude and angle)
- jet transverse energy: sum of found clustered jets
(magnitude) - 12 jet counters based on rank and position criteria
These trigger objects are calculated/extracted by the GCT
from information supplied
by the Regional Calorimeter Trigger (RCT). The original task
of the GCT was to sort
electron and jet trigger objects received from the RCT using
rank. Jets are
subdivided into and for central, forward and tau jets based
on a tau veto bit and
eta. The task of the GCT has now been extended to
perform jet cluster finding and
subsequent conversion of jet energy to rank to create
trigger objects before
performing the sort.
The electron sort operation must determine the 4 highest
rank objects from 72
candidates for both isolated and non-isolated electrons from
a significant data
volume (29Gb/s per electron type).
The jet cluster finding and subsequent sort is more
challenging because of the
larger data volume (172.8 Gb/s) and the need to share data
between processing
regions to perform cluster finding. The latter can require
data flows of a similar
magnitude to the incoming data volume depending on the
cluster algorithm used. The
baseline algorithm of a 3x3 sliding window requires
substantial data sharing, making
the system more complex. An alternative algorithm is
presented and compared to the
original.
In addition to these tasks the GCT: (a) acts as a readout
device for both itself and
the RCT by storing information until receipt of a L1A and
then sending the
information the DAQ via a SLINK64 interface (b) extracts
trigger information for the
muon system from the calorimeter data stream (c) monitors
the LHC luminosity.
The revised design is discussed, although the hardware
details are kept to a minimum
as they are presented in a separate talk at this conference.
The data processing
firmware, in particular the algorithms, data flow and
associated latency within the
revised GCT are presented.
