CLICdp New-Software Project-Meeting

Europe/Zurich
Vidyo

Vidyo

Andre Sailer (CERN)
Videoconference Rooms
CLICdp_New-Software_Project-Meeting
Name
CLICdp_New-Software_Project-Meeting
Description
CLICdp_New-Software-Project Meeting
Extension
10540573
Owner
Andre Sailer
Auto-join URL
Useful links
Phone numbers

Software Meeting Minutes
---

# 201103: Software Meeting

Agenda: https://indico.cern.ch/event/972001/

Connected: Andre, Placido, Valentin, Ulrike,
Present:

Apologies:


## MC Generators

* Some issue with using integrated processes, still takes long time when using the files
    * Solution proposed by Whizard authors: simulate one event to trigger additional calculations by Whizard for the resonance history
    * Works locally, runs much faster.
    * Issues with running on the grid: still take too long, or use too much memory, or both, for 4q2l sample
    * 2q4l samples running fine, 3% failure rate (acceptable)
    * Check with VTune (or prmon) for looking at memory usage
        * VTune: `source /cvmfs/projects.cern.ch/intelsw/psxe/linux/all-setup.sh`
        * Prmon: https://github.com/HSF/prmon
        * SystemMonitor app
    
* More sindarin files https://gitlab.cern.ch/CLICdp/DetectorSoftware/clic-whizard2-settings


## 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

## iLCDirac

* New version coming ASAP
    * DiracOS v1r11 tested
* Calice moving from LCG FileCatalog to DFC, keep v6r22 with lcg bundles, because those contain LFC bindings still

## key4hep


### CI

* https://cern.service-now.com/service-portal?id=ticket&table=u_request_fulfillment&n=RQF1650416
* iLCSoft deployment: https://gitlab.cern.ch/CLICdp/SoftwareConfigurations/iLCSoft

* Nightly build

* Working on incremental builds and deployments
    * Will be controlled from gitlab
* https://gitlab.cern.ch/key4hep/k4-deploy

* spack test -> recursion error
    * https://github.com/key4hep/k4-spack/issues/89

* Integrate with CDash, post updated packages

* Changes in spack (mesa) broke nightlies. will rebuild packages
* Good to test spack HEAD

#### Environment Script

* Overwrite PATH etc.?
    * lcg stack prepend
* Module way is slow

* `spack environment`, like the key4hep-stack bundle package

* adapt the CVMFS publisher to create the script

#### View

* WIP, also Thomas Madlener looking into it

### key4hep-spack (renamed)

* updated key4hep-stack bundle package
    * PR: https://github.com/key4hep/k4-spack/pulls
* spack namespace: k4 -> key4hep ?

### LCFIPlus inside Gaudi

* Valentin looking into it because of interest from FCC side

### podio

### EDM4hep

* builds against dev3 not working
    * TODO: Andre: try dev3 ddsim
    * TODO: dev3 builds for DD4hep disappeared

### k4FWCore

- Error on K4DataSvc

```bash
plfernan@pclcd28:~/workspace/GMP_test/build ❯ ./run gaudirun.py ../GMPWrapper/examples/import_edm4hep.py
# setting LC_ALL to "C"
# --> Including file '/home/plfernan/workspace/GMP_test/GMPWrapper/examples/import_edm4hep.py'
# WARNING: : Configurable K4DataSvc not found in module FWCore.FWCorePluginsConf
Traceback (most recent call last):
  File "/home/plfernan/workspace/gaudi_install_test/scripts/gaudirun.py", line 547, in <module>
    exec (o, g, l)
  File "<string>", line 1, in <module>
  File "/home/plfernan/workspace/gaudi_install_test/python/GaudiKernel/ProcessJobOptions.py", line 502, in importOptions                                                                                            
    _import_function_mapping[ext](optsfile)
  File "/home/plfernan/workspace/gaudi_install_test/python/GaudiKernel/ProcessJobOptions.py", line 470, in _import_python                                                                                           
    exec (code, {})
  File "/home/plfernan/workspace/GMP_test/GMPWrapper/examples/import_edm4hep.py", line 11, in <module>
    podioevent = K4DataSvc("EventDataSvc")
TypeError: 'NoneType' object is not callable
plfernan@pclcd28:~/workspace/GMP_test/build ❯
```

### k4MarlinWrapper (renamed)

* addded further test for converter
* Implement constant and "include" features used in Marlin XML
* TODO: rename the package

 

#### LCIO <-> EDM4hep conversions

* Conversion service from Gaudi
    * under investigation by Valentin
    * At the moment only used for geometry conversion
* How to do this most generically?
* How to tell which collections to import and export for given wrapper processor?

 

### spack

* upstream spack broke things via the "mesa" package
    * breaks nightly builds
    * add option to pick spack version
* PR in spack: root recipe overhaul
    * https://github.com/spack/spack/pull/19642
    * needs testing


### Gaudi

* cmake-modernization PR merged: https://gitlab.cern.ch/gaudi/Gaudi/-/merge_requests/986#note_3895786
* tried it out with k4fwcore already: https://github.com/vvolkl/K4FWCore/tree/cmake-modernization
    - working so far, but no K4FWCoreConfig.cmake generated! (policy is for users to generate their own )


## FCCee Detector Model

* Some drivers used from lcgeo moving to FCC or better DD4hep repository

## SCT Integration

* Issue with compiling DDG4 on novosibirsk cluster
   * Added: DDRec DDDetectors UtilityApps

* Add TPC xml as "Aurora" Package

## Snowmass


## AOB

 

There are minutes attached to this event. Show them.
    • 13:00 14:00
      Round the table 1h