Roundtable
Doga
How to improve track-jet and truth-jet classes. PR is almost finished and can be reviewed soon.
Guidetraining AVC and Datacenter
Stephen
Debug track finding issues in G200.
Neza
Reconstruction from truth track and truth seed. truth track works but truth seed not so well.
Alex
Working on CTD Poster. Plan to finish results this week, make poster draft on weekend, do corrections next week.
Paul
colliderML data production workflow. Disk-space problems, because 200MB/event. updating the ODD keeps it still large but under control. Not sure yet how to entangle data in root, now we need to copy everything over and over. Very similar to xAOD.
Open:
- Other signal processes, like gluon interactions.
- pile-up simulation and merging
Benjamin: how to ensure that pile-up events are pariwise intersected empty.
Luthien
Efficient programming school. Mostly parallel programming, C++ basics, openMP, thrust. Maybe too much at once, should have stuck to a single framework.
Benjamin: Had the same critisism last time.
Benjamin
Pubnote GNN4ITk needs to be finished soon, but is on track.
CTD slides need to be finished soon too.
Dimitra
Fixing some Gen3 geometry issues from Athena side. Gen2 removal in Athena already started.
SimHits are somehow out of bounds. Not sure why.
Enabled seeding from muon small wheel but not great yet.
Attila
Merging of data models in traccc.
Review on an experiment.
switching to new ntuple format saves a lot of disc space.
Paul: working on software grant proposal to integrate it into ACTS.
Carlo
Going deeper into ACTS trigger implementation. Running on truth specific chains.
Still resolution problem to fix with Jiri.
Timing: All algorithms except CKF produce times as expected. CKF reproduces old ITk reconstruction and is a lot slower.
Noemi
Seminar on tracking ITk in Krakow. It went fine, not too many people listening though. Plan to share slides with us.
Plan to prepare CTD talk.
Next week also open software community event. Thursday morning, CERN signs document with german administration. ZenDiS, STA, Red Cross.
Andreas Salzburger
Python Plugins are over, but CI does not work because of logging issue.
Geometry conversion plugin to detray. Reduced 50 candidates to 4.
Joana
Managed to rebase on new EDM. Still some segfaults but getting there. Probably some unit-conversion stuff makes problems. Annulus bounds make problems again.
Trying to traccc subspace.
PF
Understand what is going on in the efficiency losses in the central eta region.
Andreas Stefl
started work on electron-finding. new header-only implementation wit multistepper.
Code is running and more electrons can be reconstructed than with normal CKF.
Left to do: clean up and look at electrons that do not get recovered and performance is also slower.
