Skip to content

Performance improvements for the SciFiForwardTracking

This MR is based on !1985 (merged)

Requires Moore!447 (merged) (ref update)

  1. Previously we would process all ut-tracks and then reprocess the ones that weren't made into long tracks in a 2nd loop.
    This is changed we now do 1st and 2nd loop for each track before moving to the next one.

  2. Base on this, I moved some stuff around to not redo that much stuff which only needs to be done once per track.

  3. some pure C++ performance stuff.

cc @sstahl @ahennequ

Edited by Milosz Zdybal

Merge request reports