Report from DAQ Simulations Hack Days Georgia Karagiorgi & - PowerPoint PPT Presentation

Report from DAQ Simulations Hack Days Georgia Karagiorgi & Michael Baird Nov. 9, 2016

Organized by the DUNE DAQ Simulations Task Force [https://cdcvs.fnal.gov/redmine/projects/dune/wiki/DAQ_Simulations] Hack Days indico page, with links to agenda and materials: https://indico.fnal.gov/conferenceDisplay.py?confId=13041

Hack Days Format: link to slides link to slides link to slides Followed by 1.5 days of just hands on sessions.

Workshop goals: ✔ Train people in using existing DAQ sim tools and MCC files Implement zero-suppression algorithms Make some plots! ✔ Workshop turnout: • 3 people attended in person (“hands-on”) • 5 people attended remotely (mostly providing training/support) We would like to and hope to draw a bigger crowd next time! (Free coffee/pizza?!)

Tutorials: 1. Data suppression and compression SW in DUNE detector simulation and beyond, by David Adams https://indico.fnal.gov/getFile.py/access? contribId=1&resId=0&materialId=slides&confId=13041 2. DUNE MC sample generation, by Gleb Sinev https://indico.fnal.gov/getFile.py/access? contribId=3&resId=1&materialId=slides&confId=13041

Results from some (rudimentary) studies using DAQSimAna analysis module, available under georgia_daqsim_zsstudies feature branch: Ran thru DetSim using default compression scheme on simulated samples of Ar49 events. Compression scheme: Zero suppression without Huffman compression (fCompression = 2). No noise simulated! Integrating over 10 events. There are 30,720 wires (= 12 APA’s x 2560 wires each). Note: ADC waveforms written only for wires which are non-empty. Hence, 12,120 entries (not 10x30,720). Without compression, each wire waveform has 4492 samples (2.246ms).

Results from some (rudimentary) studies using DAQSimAna analysis module, available under georgia_daqsim_zsstudies feature branch: Ran thru DetSim using default compression scheme on simulated samples of Ar49 events. Compression scheme: Zero suppression without Huffman compression (fCompression = 2). No noise simulated! After compression, each wire waveform is reduced to an average of 29.75 samples.

Results from some (rudimentary) studies using DAQSimAna analysis module, available under georgia_daqsim_zsstudies feature branch: Ran thru DetSim using default compression scheme on simulated samples of Ar49 events. Compression scheme: Zero suppression without Huffman compression (fCompression = 2). No noise simulated! Compression factor = 29.75/4492 = 0.0066 A factor of ~150 reduction.

Can further study compression factor as a function of • Compression scheme • Compression parameters (e.g. threshold, number of samples to keep, etc.) Notes on factor of x150 reduction: This is just for Ar39, no noise simulation, no Huffman. Wires with no physics information are not counted! So the effective compression factor is lower: 29.75x12120 / (10x30,720x4492) = 0.000026, or ~4x10 3 reduction Target: x10 4 reduction (CDR)

Studies, continued…

A quick-start guide (so you can make these plots on your own, or contribute new ones!): cd /dune/app/users/<username> mkdir dunedaqsim cd dunedaqsim source /grid/fermiapp/products/dune/setup_dune.sh setup dunetpc v06_13_01 -q debug:e10 mrb newDev source localProducts_larsoft_v06_13_01_debug_e10/setup cd srcs mrb g dunetpc cd /dune/app/users/<username>/dunedaqsim/build* mrbsetenv mrb i –j16 mrbslp make install –j16 cd dunetpc git checkout feature/georgia_daqsim_zsstudies git pull cd dune/DAQSimAna Lar –c daqsimana_job.fcl –s <path_to_input_file>/<inputfile> You can run on any input MC file which has ben run past simwire. The file used for these studies is here: /dune/app/users/georgiak/dunedaqsim/my_prodbkgd_ar39_dune10kt_1x2x6_daq.root

Final remarks: We have some very basic tools to benchmark compression schemes. Just addressing data rates for now. Physics performance metrics will need to be defined, developed. We would greatly welcome contributions and collaboration! Thinking of having another DAQ sim hack days soon! In December? At upcoming collaboration meeting? Include light simulation studies as well? Trigger development and studies? Joint with SNB/LE?