Add basic PbPb sequence

This MR defines a new sequence hlt1_PbPb_veloSP. This sequence performs VELO, PV, and ECAL reconstruction with no GEC. The MR also introduces a heavy ion line that selects events based on total ECAL energy, number of VELO tracks, and number of PVs. Events are also selected based on whether these tracks and PVs originate from the pp or SMOG regions. This selection can be configured to handle all of the PbPb and Pb-SMOG trigger lines. The sequence runs with a throughput of about 12 kHz on a RTX 3090.

Remaining to-dos include:

• Create remaining PbPb and Pb-SMOG trigger lines
• PV reconstruction has been modified to not crash on large events. This should be fixed in a more robust way.
• Resize slice fragments to prevent Allen from hanging on large events
• Calculate total ECAL energy
• An ECAL-based lumi line should be implemented. The total ECAL energy for each event should also be saved so it can be used for centrality determination.
• Add a PbPb CI test

FYI @chwong

Also FYI @dcraik, as this concerns lumi lines and possibly monitoring.

added RTA selections labels

assigned to @thboettc, @baudurie, and @chwong

changed the description

@thboettc nice to see this! What is the reason to start with veloSP and not retina?

added 1 commit

• c6a8f23d - inital setup for UPC lines

Compare with previous version

added 1 commit

• 29046060 - Add PbPb sequence with retina decoding

Compare with previous version

@thboettc, sorry for the late answer. I don't know anything about the retina bank, but the MC sample is quite old indeed. Maybe you can try with the following sample: sim+std://MC/Upgrade/PbPb-Beam2510GeV-Upgrade-MagDown-Fix1-Epos/Sim10-Up07/24142001/XDIGI.

On another topic, as I am working on the Cal LLT encoding for Allen, I realised these LLT will only contain the transverse energy and not the total energy deposit (i.e sum of the ADC energy). While the former will be needed for the luminosity line, the latter is what we need for centrality measurement. Could we make sure that we keep this information somewhere, not only to 'label' events but throughout the reconstruction?

added 1 commit

• 2cf0e206 - Add total ECAL energy altorithm

Compare with previous version

@thboettc, thanks for the push! At the moment it is the some of ADC count that you have implemented, correct? There is a correction formula to apply which will soon be available on Moore. For the moment we can keep the sum of ADC counts.

To answer your question, we will truncate very busy/central events at the HLT2 level. We bet that we don't have to do that at the HLT1 level, but this needs to be verified ... crossing fingers

Things left to do:

• Lumi line: we can still them grom pp configuration once they are ready.
• UPC "tags" to be finalised.

These two steps could be done later, I would go with the merging and test this first configuration on PbPb data as soon as possible, if everyone agrees.

Hi @thboettc and @chwong. I admit I lost track of this MR. What is the status of the PV bug?

On my side:

• For the LLT decoding, things are on their way but I would like to test the HLT1 sequence as quickly as possible with the current SumEcal algorithm that you tested.
• For the Ecal-based Lumi lines, some work has started in lumi_reconstruction.py. Is that enough or do we need something else?

Also, there is a Pb-SMOG sample available in TestFileDB.py (Upgrade_PbAr_D0Lc_mdf). Could we test the sequence on it as well? The event rate is 300kHz if this is useful...

added 1 commit

• a38331e5 - Revert max PVs and increase event size in MDF provider

Compare with previous version

added 628 commits

• a38331e5...27e4c9fb - 619 commits from branch master
• abf13850 - Add some useful sequences for studying PbPb performance
• 31ee2311 - Create draft 0 of PbPb sequence
• c8ec7588 - Increase maximum number of reconstructed PVs
• 9167603a - Add heavy ion event line
• 21f6a677 - inital setup for UPC lines
• 4fe958ed - Add PbPb sequence with retina decoding
• 05aedc30 - Add total ECAL energy altorithm
• 0abf1d5a - Revert max PVs and increase event size in MDF provider
• c3100339 - Use memcpy_async in TotalEcalEnergy

Compare with previous version

added 2 commits

• fb31bdd7 - Remove old algorithms.py file
• 225da7df - Increase event size for MDF provider

Compare with previous version

added 2 commits

• 3c229491 - Fixed formatting
• c254bc01 - Add copyright

Compare with previous version

marked the checklist item PV reconstruction has been modified to not crash on large events. This should be fixed in a more robust way. as completed

marked the checklist item Create remaining PbPb and Pb-SMOG trigger lines as completed

added 29 commits

• c254bc01...45fc24a0 - 16 commits from branch master
• 84754d5b - Add some useful sequences for studying PbPb performance
• 1c8d45b6 - Create draft 0 of PbPb sequence
• fc5564f7 - Increase maximum number of reconstructed PVs
• b9574795 - Add heavy ion event line
• be0a82cf - inital setup for UPC lines
• 7ad9103a - Add PbPb sequence with retina decoding
• 31913b10 - Add total ECAL energy altorithm
• 8aee6b31 - Revert max PVs and increase event size in MDF provider
• a99b48ac - Use memcpy_async in TotalEcalEnergy
• db3c63d0 - Remove old algorithms.py file
• 4957440f - Increase event size for MDF provider
• abcf873a - Fixed formatting
• 8709778c - Add copyright

Compare with previous version