Updates of the new tracking tools for ITk version of the TrigFastTrackFinder, see EFTRACK-142 for more detail. No change in test output is expected.
Dmitry Emeliyanov (84d78330) at 15 Mar 17:01
implemented full road prediction algorithm, bugfix in the track fol...
... and 12880 more commits
thanks for the reminder, I've just added one more fix and made the MR ready for review
Dmitry Emeliyanov (c4b72f7e) at 13 Feb 17:28
disable tool handle if tool is not used, e.g. in non-ITk setup
The standalone mode means that the FTF is being run as an offline algorithm so that there's no external input RoI as such. When the dummy RoI is created the fullscan flag is not used because the default settings of the RoiDescriptor make little sense for ITk especially if we care about CPU timing. In https://gitlab.cern.ch/atlas/athena/-/blob/main/DetectorDescription/RoiDescriptor/Root/RoiDescriptor.cxx?ref_type=heads#L59
if ( m_fullscan ) construct( 0, -5, 5, 0, -M_PI, M_PI, 0, -s_zedWidthDefault, s_zedWidthDefault );
eta [-5,+5] is way too wide, same for the default s_zedWidthDefault which results in 450mm wide beamspot area. The original setting for zed_minus,zed_plus was indeed 180mm, but according to Elliot's studies it can be reduced down to 150mm with no loss of efficiency.
With best regards, Dmitry.
I prefer to use the "vintage" for in case I need the loop argument to address some other collection (simultaneous iteration over several collections). Unfortunately, the current C++ used in athena has no analog for Python "zip".
Dmitry Emeliyanov (6d475441) at 09 Jan 13:06
Apply 1 suggestion(s) to 1 file(s)
Dmitry Emeliyanov (9b47922e) at 09 Jan 13:05
Apply 1 suggestion(s) to 1 file(s)
yes, it will be used later when more code is added. Basically, the development plan is to have the prototype in athena, prepare/test a modified version of TrigFastTrackFinder retrieving the tool, and finally add the rest of the tool functionality.
Dmitry Emeliyanov (b4bb1419) at 08 Jan 13:56
Apply 1 suggestion(s) to 1 file(s)
Dmitry Emeliyanov (96806096) at 08 Jan 13:56
Apply 1 suggestion(s) to 1 file(s)
added a new tool prototype, development for ITk Phase II trigger tracking as described in EFTRACK-142, no changes in test output expected.
Dmitry Emeliyanov (c64e9360) at 04 Jan 18:17
added a new tool prototype, development for ITk Phase II trigger tr...
... and 9308 more commits
Dmitry Emeliyanov (adaf64f2) at 15 Dec 14:35
added checks on tool retrieval
added the option to use the fast track following tool from TrigInDetTrackFitter to facilitate high-stat tests, more details in EFTRACK-142, the option ("useTrigTrackFollowingTool") is set to false by default so that no change in tests output is expected.
Dmitry Emeliyanov (a6b35943) at 14 Dec 16:02
added the option to use the fast track following tool from TrigInDe...
... and 8997 more commits
the overall effect clearly depends on how much time is spent on combinatorial track following in FTF. With the new track following tool this time is reduced by factor of 2 for ITk geometry thanks to many Pixel layers in ITk. For the current ATLAS geometry, it needs to be tested.
many thanks for pointing this out! It should be fixed with the latest update.
Dmitry Emeliyanov (cd7d86c5) at 06 Dec 17:19
fixed compilation failure due to a change in offline ctors
Hi Mark, it could be adapted to the current geometry with some minor changes. For instance, module shape type would be different for SCT endcaps as Annulus-type is introduced in ITk and basically replaces Trapezoids.