Prepare downstream tracking for 2024 data-taking test
Tune the HLT1 Downstream setup with run 301222 (μ=4.4, standard UT setup).
Output rate in mu=4.4:
Hlt1DownstreamKsToPiPi: 784/960000, ( 24.50 +/- 0.87) kHz
Hlt1DownstreamLambdaToPPi: 577/960000, ( 18.03 +/- 0.75) kHz
Hlt1DownstreamPromptKsToPiPi: 2/960000, ( 0.06 +/- 0.04) kHz
Hlt1DownstreamPromptLambdaToPPi: 1/960000, ( 0.03 +/- 0.03) kHzInclusive(Hlt1Downstream.*): 36.65 kHz
Output rate in mu=5.3:
Hlt1DownstreamKsToPiPi: 594/600000, ( 29.70 +/- 1.22) kHz
Hlt1DownstreamLambdaToPPi: 464/600000, ( 23.20 +/- 1.08) kHz
Hlt1DownstreamPromptKsToPiPi: 0/600000, ( 0.00 +/- 0.00) kHz
Hlt1DownstreamPromptLambdaToPPi: 0/600000, ( 0.00 +/- 0.00) kHzInclusive(Hlt1Downstream.*): 44.55 kHz
Note the output rate does scale linearly!
Changes in the code:
- Added overflow counter for downstream tracking for monitoring
- Extended the max container size of downstream tracking from 2048 to 4096
Note: This should be reviewed by WP2 instead of WP3, as the only changes relate to reconstruction.
Update (Sep 27)
The merge request !1766 (merged) has been merged into this branch: the new line Hlt1DownstreamD2Kshh has been added. To keep the total HLT1 downstream line output rate under 50 kHz, we adjusted the Hlt1DownstreamKsToPiPi threshold from 0.5 to 0.55, which reduced the rate from 29.70 kHz to 23.20 kHz.
Updated output rate at mu=5.3 (Run 306411)
Hlt1DownstreamKsToPiPi: 464/600000, ( 23.20 +/- 1.08) kHz
Hlt1DownstreamLambdaToPPi: 465/600000, ( 23.25 +/- 1.08) kHz
Hlt1DownstreamPromptKsToPiPi: 0/600000, ( 0.00 +/- 0.00) kHz
Hlt1DownstreamPromptLambdaToPPi: 0/600000, ( 0.00 +/- 0.00) kHz
Hlt1DownstreamD2Kshh: 33/600000, ( 1.65 +/- 0.29) kHzIn order to keep the git history clearer, this MR disables the final squash, and the commits should be cleaned up using git rebase + fixup
FIY: @adeoyang @dovombru @mveghel @gligorov @decianm @cagapopo