Level-3 Trigger Strategy

From GlueXWiki
Jump to: navigation, search

This page will collect the currently planned running times/intensities for GlueX and the expected event rates. A set of requirements for the Level-3 trigger will be collected based on this information to document the design and implementation strategy.

Phase II and III Planning

After the initial Phase I running period, designated for detector commissioning, the first "physics" running periods are Phases II and III, which both have a nominal photon flux of 10^{7} in the coherent peak (8.4-9 GeV). The expected Level-1 trigger rate for these periods is expected to be at or below 20 kHz, and therefore it is possible to write the entire dataset to tape with a minimum-bias type L1 trigger.

If the Level-3 trigger is demonstrated to provide sufficient efficiency with a useful rate reduction factor before the end of the Phase II or III running periods, it could be used to reduce the event rate written to tape before Phase IV running. This should save on storage space as well as reconstruction/analysis turnaround time which could reduce some CPU cost.

The goal in Phase II and III is to commission the Level-3 trigger for future running periods by:

  1. Running a Level-3 trigger algorithm in pass-through mode on a subset of nodes in the monitoring farm.
  2. Tagging events in the data stream which satisfy the algorithm requirements to allow studies offline of the:
    • Trigger efficiency in the data with full offline reconstruction
    • Rate reduction factor to verify it meets the requirements for future running

Phase IV Planning

This is the first phase of GlueX running where the Level-3 trigger is required to reduce the event rate beyond the reduction provided in hardware by the Level-1 trigger. The nominal photon flux for this phase is 5x10^{7} in the coherent peak, which is expected to result in a Level-1 trigger output rate of around 100 kHz based on previous studies. Given the event size and DAQ bandwidth expectations in Table 2 (below), the event rate written to tape is limited to 20 kHz.

The goal in Phase IV is to reduce the Level-1 trigger rate to 20 kHz (factor of 5 reduction), while accepting all reconstructable exotic-meson candidates produced in the coherent peak.

Full Intensity Planning

The GlueX design goal was for a photon flux of 10^{8} in the coherent peak, but the optimal photon flux may depend on many factors including beam conditions, pileup, etc. An additional complication of this full intensity running is that it may not be possible to record all the hadronic events in the coherent peak in the allowed 20 kHz of DAQ to tape bandwidth, given realistic limitations on the L3 algorithm. Therefore, a more selective sample of the highest priority physics channels would need to be identified, such as identifying strangeness candidates via displaced vertices or an upgraded forward PID detector. This will require further studies in the future, but is not a high priority for the Trigger Working Group at the current time.

Assumptions for Level-3 Trigger Planning

Some assumptions are necessary to plan for the requirements of the Level-3 trigger based on the expected running conditions of the experiment. Below are two tables with some assumptions for the data taking conditions for the GlueX experiment.

Table 1: Planned GlueX Running Periods (from PAC 40 Proposal)
Phase I Phase II Phase III Phase IV Full Intensity
Approved PAC days 30 30 60 200 0
Minimum electron energy (GeV) 10 11 12 12 12
Photon flux in coherent peak (photon/s) 10^{6} 10^{7} 10^{7} 5x10^{7} 10^{8}
Expected total hadronic event rate (kHz)  ??  ?? 36 180 360
Expected hadronic event rate in coherent peak (kHz)  ??  ?? 1.6 8 16
Level-1 trigger rate (kHz) 2 20 20 100 200


Table 2: Data rate assumptions for GlueX
Event size 15 kB
DAQ bandwidth to write events to tape 300 MB/s