Add minimum momentum cut in hybrid seeding
Without cut:
SeedTrackChecker_6f38a9cb INFO Results
SeedTrackChecker_6f38a9cb INFO **** Seed 9558 tracks including 479 ghosts [ 5.01 %], Event average 4.82 % ****
SeedTrackChecker_6f38a9cb INFO 01_hasT : 6497 from 7781 [ 83.50 %] 0 clones [ 0.00 %], purity: 99.56 %, hitEff: 98.00 %
SeedTrackChecker_6f38a9cb INFO 02_long : 4642 from 4960 [ 93.59 %] 0 clones [ 0.00 %], purity: 99.70 %, hitEff: 98.60 %
SeedTrackChecker_6f38a9cb INFO 03_long_P>5GeV : 3291 from 3420 [ 96.23 %] 0 clones [ 0.00 %], purity: 99.67 %, hitEff: 99.09 %
With cut at 4500:
SeedTrackChecker_6f38a9cb INFO Results
SeedTrackChecker_6f38a9cb INFO **** Seed 6894 tracks including 284 ghosts [ 4.12 %], Event average 3.88 % ****
SeedTrackChecker_6f38a9cb INFO 01_hasT : 4998 from 7781 [ 64.23 %] 0 clones [ 0.00 %], purity: 99.55 %, hitEff: 98.67 %
SeedTrackChecker_6f38a9cb INFO 02_long : 3816 from 4960 [ 76.94 %] 0 clones [ 0.00 %], purity: 99.69 %, hitEff: 99.04 %
SeedTrackChecker_6f38a9cb INFO 03_long_P>5GeV : 3292 from 3420 [ 96.26 %] 0 clones [ 0.00 %], purity: 99.67 %, hitEff: 99.09 %
I will be very curious to see the speedup. The cut is done at the earliest possible level and so the speedup is probably much, much better than nTracks(cut)/nTracks(without cut). I am more expecting a factor 4 or so, and if needed I can provide an even faster version (factor 2 or more), but this will start actually cutting slightly on performance.
Edited by Louis Henry