CLICdp New-Software Project-Meeting

Europe/Zurich
1/1-025 (CERN)

1/1-025

CERN

6
Show room on map
Andre Sailer (CERN)
Videoconference
CLICdp New-Software Project-Meeting
Zoom Meeting ID
95706486793
Host
Andre Sailer
Useful links
Join via phone
Zoom URL

Software Meeting Minutes
---

# 210804: Software Meeting

Agenda: https://indico.cern.ch/event/1064845/

Connected: Raffaela
Present: Andre, Placido, Valentin

Apologies:

## Proceedings/Talks/etc.

 

## Summer Student project

 - https://codimd.web.cern.ch/omk0BJcXTqiITllS-JGlLQ?edit#
 - [ ] fetch downstream changes from Muon Collider fork of DDDigiPlanarProcessor

## MC Generators

## CLICDet Delphes Card

## Simulation

* Issue with simulation of muons with fixed pT?
    * observed different duplicate vs phi using different muon simulation runs
    * Maybe different input (not pt, but energy)?
    * Different Theta range, 0-90 degrees before, 0-180 degrees now

## Reconstruction

### Tracking

### Particle Flow

### Digitizer

### Event Display

### FlavourTagging

### Long-Lived Particle

### dE/dX

### LCGeo Test

## Infrastructure

### Sanitizers/Compilers

Placido: Investigating Santizers for the MarlinWrapper, but cannot compile due to Kernel issue
https://github.com/google/sanitizers/issues/856
```
plfernan@pclcd28:~/workspace/clang_k4mw/k4MarlinWrapper/cmake ❯ make         
[  7%] Linking CXX shared module libLcio2EDM4hep.so
==1077==Shadow memory range interleaves with an existing memory mapping. ASan cannot proceed correctly. ABORTING.
==1077==ASan shadow was supposed to be located in the [0x00007fff7000-0x10007fff7fff] range.                              
==1077==This might be related to ELF_ET_DYN_BASE change in Linux 4.12.  
==1077==See https://github.com/google/sanitizers/issues/856 for possible workarounds.
==1077==Process memory map follows:
        0x000000400000-0x000000409000   /cvmfs/sw.hsf.org/spackages/linux-centos7-broadwell/gcc-8.3.0/gaudi-35.0-wcrxdxolxgxqwtx3w3rgnzjrbkoqlh3j/bin/listcomponents
        0x000000608000-0x000000609000   /cvmfs/sw.hsf.org/spackages/linux-centos7-broadwell/gcc-8.3.0/gaudi-35.0-wcrxdxolxgxqwtx3w3rgnzjrbkoqlh3j/bin/listcomponents
        0x000000609000-0x00000060a000   /cvmfs/sw.hsf.org/spackages/linux-centos7-broadwell/gcc-8.3.0/gaudi-35.0-wcrxdxolxgxqwtx3w3rgnzjrbkoqlh3j/bin/listcomponents
...

```

## iLCDirac

* FCC VO
    * EOS integrated
        * Cannot be in more than one VO
    ```
    source /cvmfs/clicdp.cern.ch/DIRAC/bashrc
    dirac-proxy-init -g fcc_user
    XrdSecPROTOCOL=gsi,unix eos whoami
    ```
    ilc001
    ```
    Virtual Identity: uid=56064 (99,56064) gid=1018 (99,1018) [authz:gsi]   host=lxplus728.cern.ch domain=cern.ch geo-location=0513
    ```
    * Cannot run key4hep stack with Broadwell on cern HTCondor

```
/bin/sh: line 1:   555 Illegal instruction     (core dumped) ./DelphesPythia8_EDM4HEP ./edm4hep-delphes-lhe-examples/simdelphes/delphes_card_IDEA.tcl ./edm4hep-delphes-lhe-examples/simdelphes/edm4hep_output_config.tcl ./edm4hep-delphes-lhe-examples/cards/Pythia_LHEinput.cmd edm4hep_test_output.root
```

* add key4hep-stack to config  (gaudiapp)
    * k4run executable (modifiable)
    * Test Clic Reco With GaudiWrapper inside iLCDirac
        * InputFiles
            * --EventDataSvc
            * --LcioInput
        * RandomSeed
            * Not fully implemented in vanilla Gaudi
            * Should implement something in Gaudi like the randomseeder in Marlin
        * OutputFile
            * PodioOutput
        * CompactFile
            * -->GeoSvc (to come to k4Sim)
        * NumberOfEvents
            * -n , --ApplicationMgr.EvtMax

## Key4hep

* build with gentoo-prefix almost working, except some packages
    * header, include things
    * whizard2 tricky

### Gaudi

* Writing guide for upgrading to new gaudi cmake
    * updated documentation for gaudi, only needs deployment now
    * https://gitlab.cern.ch/gaudi/Gaudi/-/blob/4b5340fbc9afa82ab1434721870d40665bf4df54/docs/source/Developers/cmake.rst
    * https://gitlab.cern.ch/gaudi/Gaudi/-/merge_requests/1152

    * Doc: gaudi.web.cern.ch
    
* Documentation for functional:
    * https://lhcb.github.io/DevelopKit/

* Make a service optional/disable in config file: Ticket with Gaudi
    * https://gitlab.cern.ch/gaudi/Gaudi/-/issues/178

### CI

* CVMFS deploy from gitlab: https://cern.service-now.com/service-portal?id=ticket&table=u_request_fulfillment&n=RQF1650416
    * --> Need to re-open the ticket (Valentin)
    * gitlab-runner  was missing permissions
        * New ticket: https://cern.service-now.com/service-portal?id=ticket&table=u_request_fulfillment&n=RQF1747119
        * IT will fix the gitlab integration on another repo and then replay the steps on sw.hsf{-nightlies}.org


### Environment Script

### View

### key4hep-spack/spack

### podio

### EDM4hep

* Need TrackerHitPlane class in EDM4hep

* UserClasses, discuss in podio/edm4hep meeting if direct access to underlying pods is possible or desirable

* Test writing sample event data:
    * Incomplete, e.g., TrackCollection not written
    * Would also be useful for testing EDM4hep <-> LCIO Conversion

### k4FWCore

* getTypeOfCollection from NameOfCollection
    * Can already get type from collection from base class


### k4MarlinWrapper

* LCRelation troubles because of mismatch between EDM4hep and LCIO way of doing relations

* geoSvc to replace MarlinDD4hep
    * Need to set the GlobalTrackerReadoutEncodingStringID

 


### spack

Issue

Opened PRs for a bunch of improvements:
* [ ] fixes to fetching fastjet https://github.com/spack/spack/pull/20064
* [ ] special "commit" version for nightlies https://github.com/spack/spack/pull/20065

* Compiling everything with clang
    * Issue with "base" packages, mesa

## FCCee Detector Model

## SCT Integration

### Using the lcgeo/ddsim/Marlin to run simulation and reconstruction

* Provide set up and instructions for running
    * Do this now, use key4hep setup.sh from cvmfs
    * run ddsim only with TPC (other detectors probably need geometry from aurora, which we cannot access)
        * DDSim on cluster doesn't work because DD4hep too old and python mismatch
    * run Marlin with reconstruction from https://gitlab.cern.ch/sailer/SCT_TPCReco/
    * run wrapped reconstruction with LCIO until conversion from SimTrackerHit is available as well
 
* Integrating TPCSD sensitive into Aurora not so straight forward
    * They directly use G4VSensitiveDetector instead of dd4hep::Geant4Action types.
    * Also event data model differences need to be understood

* WIP: Add TPC xml as "Aurora" Package
* Part of simulation, but now need to create the hits.

 

## Snowmass

## AOB

### Next Meeting

August 10, 2021, 13:00

 

 

There are minutes attached to this event. Show them.
    • 1:00 PM 2:00 PM
      Round the table 1h