cms-L1TK / cmssw

Fork of CMSSW where improvements to L1 tracking code are developed.
http://cms-sw.github.io/
Apache License 2.0
4 stars 5 forks source link

Cbrown kfout #93

Closed Chriisbrown closed 3 years ago

Chriisbrown commented 3 years ago

PR description:

Adds the KFout emulator into the new KF development branch. This module acts to emulate the FW implementation of the TTTrack packing and routing to assign tracks to the correct links. It includes the approximate chi2 calculations including additional values in the Producer_setup config file. It also includes minor additions to the AnalyzerDemonstrator module in order to allow for just tracks to be put into pattern files without stub information.

PR validation:

This has been run comparing emulator output to expected SW output as well as FW output, it does not include a clock accurate emulation of the router logic used in FW due to complexity so is not a clock accurate emulation

tschuh commented 3 years ago

In order to validate s/w and f/w we will need a clock accurate emulation. Is that something you plan to add in near future?

tschuh commented 3 years ago

Please merge with tschuh12. The recommended way works as follows:

git cms-checkout-topic -u cms-L1TK:tschuh12
git cms-rebase-topic -u cms-L1TK:cbrown-kfout
git push -f cmsL1TK cbrown-kfout

where I assumed that you called the remote cms-L1TK.

tschuh commented 3 years ago

When I run L1Trigger/TrackFindingTracklet/test/HybridTracksNewKF_cfg.py I see debug prints and a Map:at std::exception.

Chriisbrown commented 3 years ago

In order to validate s/w and f/w we will need a clock accurate emulation. Is that something you plan to add in near future?

This will be done, but unsure of the timescale there are a few bits that need sorting out, we can merge the current changes to the kfout now or wait until I have got the clock accuracy correct

Chriisbrown commented 3 years ago

When I run L1Trigger/TrackFindingTracklet/test/HybridTracksNewKF_cfg.py I see debug prints and a Map:at std::exception.

Issue with accessing the ttrackrefMap has been solved, Map:at exceptions no longer occur