REQUEST FOR PRODUCTION RELEASES:
To request your PR to be included in production software, please add the corresponding labels called "async-" to your PR. Add the labels directly (if you have the permissions) or add a comment of the form (note that labels are separated by a ",")

+async-label <label1>, <label2>, !<label3> ...

This will add <label1> and <label2> and removes <label3>.

The following labels are available
async-2023-pbpb-apass4
async-2023-pp-apass4
async-2024-pp-apass1
async-2022-pp-apass7
async-2024-pp-cpass0
async-2024-PbPb-apass1
async-2024-ppRef-apass1
async-2024-PbPb-apass2
async-2023-PbPb-apass5

Error while checking build/O2/fullCI_slc9 for 1e0a4df at 2025-07-25 17:34:

++ cp /sw/slc9_x86-64/O2/14544-slc9_x86-64-local1/compile_commands.json .
+++ python3 -c 'import json, os; print(os.path.commonprefix([x["file"] for x in json.loads(open("compile_commands.json").read()) if "sw/BUILD" not in x["file"] and "G__" not in x["file"] and x["file"].endswith(".cxx")]))'
++ O2_SRC=/sw/SOURCES/O2/14544-slc9_x86-64/0/
++ [[ -e /sw/SOURCES/O2/14544-slc9_x86-64/0//CMakeLists.txt ]]

Full log here.

lgtm

Hi @shahor02 - could we please merge this? It changes nothing in practice (due to backwards-compatible defaults) but adds a settable max XIU for track pre-propagation at AO2D creation time that, following a discussion in the last Physics Board meeting, we should explore for deployment and saving resources. We should also discuss this in one of the next software meetings after further testing and setup with the DPG. Thank you!

Hi @ddobrigk, @f3sch

I was not present at the last PB. Just to be sure: this PR effectively makes only tracks with ITS IB contribution allowed to propagate to the PV instead of those with any ITS contribution. Is this what you want?

Hi @ddobrigk, @f3sch

I was not present at the last PB. Just to be sure: this PR effectively makes only tracks with ITS IB contribution allowed to propagate to the PV instead of those with any ITS contribution. Is this what you want?

Yes, that's effectively the new default value. The point is that in the vast majority of analyses, ITS IB contributions are required for primary tracks, and thus the track propagation task is used anyways to propagate only tracks with XIU < 5cm as stored in the AO2Ds. With this default value of 5cm, we do at AO2D creation time the majority (>95%) of the track propagation that gets done at analysis time (modulo the small O(1%) tracks used by the svertexer), saving a lot of CPU. Tracks that only have OB contributions are kept where they are in the new default value.

However, the PR also makes this value settable in case there's the desire to do so. Furthermore, the propagate-tracks option is anyway false everywhere right now, so the current behaviour remains unchanged in case nothing is changed in the DPG scripts - which we will take care to do only very carefully.