- Moved HitFactory2D class to dedicated file.
- Parallelized the entire chain "cluster finding -> hit finding -> hit merging"
  by row index. Merging is done in parallel between pairs of neighboring rows,
  using even-odd partitioning (fully supported for 2D, interface supplied for 1D).
- Cleaned up TRD reconstruction classes and fixed several warnings.

2) the mcbm_beam_2024_05_08_nickel setup is added to the CbmMcbmUtils (for RunID > 2917)

Sergei Zharko authored 1 year ago and Sergei Zharko committed 1 year ago

Change of parameter repository hash to support MR CbmSoft/cbmroot_parameter!180

Sergei Zharko authored 1 year ago and Sergey Gorbunov committed 1 year ago

- tracking interfaces for online STS and TRD
- installation rule for the macro/beamtime/mcbm2024 directory

- Add High-water Mark setting on the ZMQ output socket with default value of 1
- Add binary option to set the HWM + help info

- introduce a `ReadoutPars` struct similar to the one used on `tof` to add support for multiple unpack configs
- add new detector as "fake" back-side of the existing diamond detector to avoid need a new geometry
- switch readout parameters based on run ID, switch point hard-coded to run `2941` of 02/05/2024
- add needed hard-coded use of the 2022 version in the "algo offline" tasks (matches TOF/STS choice)

- Fix support for TOF bucarest boxes
- Change the type of the last RPC in `ReadoutPars_mCBM2024.cxx`
- Add log printout of the TOF mapping at debug level

Significantly improves performance for high combinatorics. Especially
important for the new Module 0 in mSTS 2024.

Pierre-Alain Loizeau authored 1 year ago and Volker Friese committed 1 year ago

- skip subsubevents with corrupt data (wrong order of special words or missing ones)
- add more errors counters
- debug logs when corrupt data found
- handle data with less words than normal but proper header info
- recover the last TimeData if only the Trailer is missing
- in ProcessTimeDataWord, fix return to avoid out of bounds access