Rongkun Wang's sandbox for storing HEP infos.
Administrative Memo
USTC Memo
Swiss plate to green plate
To get things published
Quick Memo
This is an assorted notes on many aspects I faced in HEP, continuously building up. Knowledge and purely technical guides are mixed up.
ATLAS Induction Day 2017
SeriesOfFacts
Physics
- Physics of Generating events
- Pileup: Some official twiki to introduce Pileup/Pileup reweighting. Pileup, an unavoidable phenomena that occurred when we're running with high luminosity. In order to increase the luminosity, more beam particles are squeezed into the same space, therefore increasing the chance of interaction. With this, we also have unwanted non-deep inelastic interactions that produced pileup jets. The rate of those kind of process can be described by number of vertices(which is used in run 1), or more precisely the average interaction per crossing <μ>.
Detector related Knowledges
Pre-Analysis Preparation
Generator related technical things
I have modified and installed a <a data-wikiword="TopDrawer" href="https://gitlab.cern.ch/rowang/TopDrawer" target="_blank">
TopDrawer </a> to visualize .top files generated after generators like powheg.
do
td
after
source /afs/atlas.umich.edu/home/rkwang/public/sw/setup_TopDrawer.sh
Object Reconstruction
Physics Analysis
batch kill panda jobs
Pbook show allows wildcard, but kill doesn't (isn't this silly?)
So you can direct the print out from show to a string, and then do some formatting to pass them to kill(), in a loop. Modify the matching_string to your wildcard string
import sys
# python 2
from cStringIO import StringIO
# python3
# from io import StringIO
matching_string = "user.rowang.WVZ_v1.2.364*"
old_stdout = sys.stdout; sys.stdout = mystdout = StringIO(); show(taskname = matching_string); sys.stdout = old_stdout
# reviewing your output ID parent
for i in mystdout.getvalue().split("\n")[2:-1]: print i.split()[0], i.split()[4];
# actual delete!
for i in mystdout.getvalue().split("\n")[2:-1]: kill(int(i.split()[0]));
ATLAS Rivet &
HepData guidline:
https://twiki.cern.ch/twiki/bin/viewauth/AtlasProtected/HepDataRivetSubmission
To know weightName in ATLAS:
checkMetaSG.py
EVNT to TRUTH1:
## truth ntuple
export ATLAS_LOCAL_ROOT_BASE=/cvmfs/atlas.cern.ch/repo/ATLASLocalRootBase
source ${ATLAS_LOCAL_ROOT_BASE}/user/atlasLocalSetup.sh
source $AtlasSetup/scripts/asetup.sh 21.2.55.0,AthDerivation,here
Reco_tf.py --inputEVNTFile $EVNT --outputDAODFile $DAOD_TRUTH1 --reductionConf TRUTH1
Computing
- Root: Experience that I had dealing with ROOT.
- GitLab: tutorial By Will Buttinger: Link. * How
to write readme.md plain text syntax
- How to prevent disconnected by lxplus when you idle for a few minutes? See this NAT related article
. 120 second works for me.
- Where to save files on lxplus? EOS! How to read eos on a non-fusemount machine: ATLASStorageAtCERN
- Condor: on occasion, the condor schedd you're allocated to becomes busy, you can change the env var, e.g.
- _CONDOR_SCHEDD_HOST=bigbird15.cern.ch _CONDOR_CREDD_HOST=bigbird15.cern.ch
- Host related issue: SetupServer
- X11 forwarding not working sometimes!
- OS installation:
ATLAS Duty
View your (Operation Task Point)OTP at
GLANCE. You bear the duty for your institute to earn OTP, and also to earn you a chance to give conference talks.
Muon Offline Data Quality (Class 2)
Shift:
Expert: MDT Operation Oncall
NSW-sTGC related
Workflow
- Connect all the frontend boards correctly.
- Load your json/miniDaq config files to FELIX GUI, tune the parameter(according to connection, which felix server, etc).
- Save the json file and OpcServer xml file on a place felix server can access.
- Set up felix server and run felixcore and OpcServer.
- Open partition, click Initialize->Config->Run, to take data, and then Stop to stop taking data.
- Analyze the binary data ended with .data. (decode into TTree see below).
Hand-on manuals
B180 felix1: Step-by-step hands-on
twiki from Roy . (obsolete)
191 sTGC: operation for sTGC
Technicalities
Related concepts and manual:
- FELIX is like a big data router, and running it doesn't rely on any other thing. See this FELIX manual for setting up and running FELIX. List of Felix server:
- um-felix1.cern.ch
- um-felix2.cern.ch
- The FELIX GUI is a stand-alone application that can generate board configuration, opc configuration, visualize and modify parameters in a better way than json. You can run it in principle on any machine. Or choose not to use it.
- The partition is the very same thing used to control ATLAS data taking. The config segment relis on FELIX and OpcServer to send HDLC frame, software ReadOut Driver(swROD) relies on FELIX for data readout. See the manual here as how to run partition with three segments swROD(ROx), TTC, Config. And also how to decode the data into a TTree. List of working swROD server:
- um-felix1-swrod.cern.ch
- um-felix2-swrod.cern.ch
- um-felix2-swrod-supermicro.cern.ch
- TTC server:
- um-felix1-spc.cern.ch
- um-felix2-spc.cern.ch
- Here is a very nice summary from our neighbor MicroMegas (MM).
- How to use UaExpert on our FELIX machines: `setupUaExpert; uaexpert`
- Server related issues are summarized here: SetupServer. See also Computing section.
- Open points:
- Scalability: it scalse well to one wedge/12+12 boards. one thing we learned is GBTx phase need to be saved and tuned after ROC is configured.
- Stability: how long can the system run? Cooling issue: observed crashes, power off to cool down.
- swROD production machine cannot run partition. - still debugging
Baseline and Trimmer Scan
cd /afs/cern.ch/work/n/nswdaq/public/sTGC_quick_and_dirty_baselines
source setup.sh
NSWConfiguration/dev/stgc_threshold_calib.sh
The output will be saved in
http://cern.ch/stgc-trimmer/
See also
https://twiki.cern.ch/twiki/bin/viewauth/Atlas/BB5DaqMM for full instruction
VMM ID to VMM capture mapping in swROD. Notice the mapping is symmetric. See also ATLNSWDAQ-74
0<-->2
1<-->3
2<-->0
3<-->1
4<-->5
5<-->4
6<-->6
7<-->7
Analysis
WVZ Analysis
MyOwnWVZDraftTwiki
HSG2 (HZZ)
H4l
- Fiducial/Differential
- Fiducial: xs in each channel, and sum, combined. Also Total xs.
- Differential: njets, pt, unfolding with bin-by-bin.
- Coupling
- STXS Intro
: Truth bin: stage0 with ggF(+bbH)/VBF/VH(+ggZH)/ttH(+tH) and reduced stage1 STXS production mode(10 categories). Reco bin: 11 categories.
- In the first round of run 2 analysis. We also had 9 reco categories.
- Conf Note 2018: 2 additional categories are added in reco level for: split ttH and added 0jet high pt4l for VHLep
- STXS POI and corresponding fiducial definition: Twiki. Acceptance calculation
- Documents
Unfolding of Inclusive M4l
Aiming to unfold m4l distribution inclusive of number of jets. We're also intereseted in unfolding 2D distribution of m4l.
- Twiki
- 2018 Paper draft cds
: supporting note in svn, in cds
- Accuracy of simulation:
- qqZZ Sherpa 2.2.2 NLO for 0,1-jets, LO for 2,3-jets. (QCD)
- Reweighted to NLO EW as func of m4l
- ZZ+2j HOEWK (VBF-like process) are produced by Sherpa 2.1.1.
- ggZZ Sherpa 2.2.2 Generation is performed at leading order+NLO (corresponding to NNLO QCD)
- Reweighted over m4l to NLO with MCFM samples generated at LO and reweighted to LO with m4l where NLO/LO k-factor is given by theorists.
- ggH and VH: Powheg NNLOPS + Pythia8
- ttH: MadGraph5_aMC@NLO + Herwig++
Standard Model
VBS ZZ
Twiki
minitree location: /afs/cern.ch/work/b/bili/public/VBS/MiniTree/
Exotics
H-->ZdZd-->4l
Search for a light BSM model in Higgs decay channel. Run 2 we merge pseudo-scalar result into our paper.
--
RongkunWang - 2019-05-03