Skip to content

L1TriggerByteStreamDecoderAlg: Reorganise ROB request handling

Rafal Bielski requested to merge rbielski/athena:improve-l1-bs-decoder into master

Reorganise ROB requests in L1TriggerByteStreamDecoderAlg with two changes:

  • Batch requests from all decoder tools into one request to benefit from downstream parallelisation (in DCM) and reduced network overhead. Then for each tool, filter out only the requested ROBs for that tool.
  • If any of the requested ROBs are missing, print an ERROR and return FAILURE. This is required to catch missing L1 data which would otherwise be silently missed in HLT seeding. Potentially would catch issues like described in ATR-23697.

cc @wiedenma

Also relevant Jira: ATR-20745

Merge request reports