1 / 28

PhenoGrid Status

Durham University. PhenoGrid Status. Peter Richardson. PhenoGrid. PhenoGrid is a new project which started with GridPP2. The project has two main aims Enabling U.K. phenomenologists to make use of the Grid.

elga
Télécharger la présentation

PhenoGrid Status

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Durham University PhenoGrid Status Peter Richardson GridPP, Durham 5th July

  2. PhenoGrid • PhenoGrid is a new project which started with GridPP2. • The project has two main aims • Enabling U.K. phenomenologists to make use of the Grid. • Developing simulation software, mainly Monte Carlo event generators, for use on the Grid. GridPP, Durham 5th July

  3. PhenoGrid • PhenoGrid is a collaboration of theoretical physicists. • It involves all interested U.K. phenomenologists. • While many of us are interested in computing, the computer is a tool to allow us to solve physics problems. • If it is too complicated, or the documentation is inadequate these people will not use the Grid. GridPP, Durham 5th July

  4. PhenoGrid VO • Due to the short timescale between the award and start of the grant it proved impossible to start before January. • One postdoc David Grellscheid started mid-Jan. • Since then much of the work has been aimed at allowing PhenoGrid members to submit jobs. • This has gone VERY slowly due to factors beyond our control. • Before any jobs can be submitted we need to set up a VO. GridPP, Durham 5th July

  5. PhenoGrid VO • Rutherford lab agreed to host the VO server. • The test server was set in Feb/March. • Due to incompatibilities between VOMS and the resource broker couldn’t submit jobs. • Solution: wait for the next upgrade, delayed everything two months. GridPP, Durham 5th July

  6. PhenoGrid VO • After this we finally got to the point we can register users and try to submit jobs. • Database corruption problem. Not worth fixing before the next upgrade. • Solution: wait for the next release, delays everything another two months delay? • One test job submitted but due to problems not via the resource broker. GridPP, Durham 5th July

  7. PhenoGrid VO • I had hoped thatthings had improved since I first refused to use the Grid three years ago as it wasn’t worth the trouble. • It doesn’t seem to have fast enough. • If things don’t get better it will be hard to persuade theorists to use the Grid. GridPP, Durham 5th July

  8. Monte Carlo Event Generators • Monte Carlo Event Generators are simulations of what happens when particles collide. • As such they are the starting point of the LHC simulation chain. • The results are then feed into simulations of the detector. • As such they are a vital part of LHC computing. GridPP, Durham 5th July

  9. A Monte Carlo Event Hard Perturbative scattering: Usually calculated at leading order in QCD, electroweak theory or some BSM model. Modelling of the soft underlying event Multiple perturbative scattering. For some processes the NLO corrections are now included. Perturbative Decays calculated in QCD, EW or some BSM theory. Initial and Final State parton showers resum the large QCD logs. Finally the unstable hadrons are decayed. Non-perturbative modelling of the hadronization process. GridPP, Durham 5th July

  10. Monte Carlo Event Generators • They have been chronically under funded. • Despite the vast manpower devoted to GEANT4 and individual detector simulations less than 10 man years has been spent on event generators. • This stupid, after all if we feed garbage in then all the rest of the effort is wasted. GridPP, Durham 5th July

  11. Herwig++ Herwig++ collaboration: S. Gieseke, M. Gigg, D. Grellscheid, K. Hamilton, A. Ribon, P. Richardson, M. Seymour, P. Stephens, B. Webber • The Herwig++ project is a major effort to produce a new event generator in C++ based on the physics models in the successful Herwig program. • There is less manpower than it seems ~ 3 FTE. GridPP, Durham 5th July

  12. Herwig++ • This project was needed for a number of reasons • The existing FORTRAN was becoming impossible to maintain. • Many new physics ideas could not be included due to the code structure. • Everyone else was moving to C++ GridPP, Durham 5th July

  13. Herwig++ • Herwig++ is based on the Toolkit for high energy Physics Event Generator ThePEG framework. • This a common framework together with the Lund family of simulations. • Most of the PhenoGrid deliverables are related to the Herwig++ project and this underlying framework. GridPP, Durham 5th July

  14. Herwig++ • For a theoretical project Herwig++ is a large collaboration. • We have weekly phone meetings to manage the various aspects of the project. • Six monthly collaboration meetings to keep track of the overall state of the project. GridPP, Durham 5th July

  15. Herwig++ • The Herwig++ project still needs a number of important features for simulating hadron-hadron collisions • Initial-State Shower • Hadron Decays • Underlying event • In order to deliver some of these we cannot use students or normally funded postdocs. • It would end their careers before they start. GridPP, Durham 5th July

  16. Initial State Shower GridPP, Durham 5th July

  17. Hadron Decays GridPP, Durham 5th July

  18. Herwig++ • The GridPP deliverables • Improved documentation • Grid running • Inclusion in the LCG GENSER library • Interfacing to LHAPDF • Interfacing to EvtGEN are due to their nature less important • This means we can not justify using unfunded effort for many of the deliverables. GridPP, Durham 5th July

  19. Herwig++ Deliverables • The first set of deliverables relating to improving Herwig++ documentation and familiarisation of postdoc with the program have been met. • The documentation has been a major problem with some theorists unable to use the code because of it. GridPP, Durham 5th July

  20. Hadron Decays GridPP, Durham 5th July

  21. Documentation GridPP, Durham 5th July

  22. Near Future • The next deliverable relates to Grid running of the program. • We have made major changes to automise the installation to make grid running and inclusion in the LCG GENSER library easier. • Can’t do anything else until we can submit jobs. • This means that this will be at least another 3 months before completion. GridPP, Durham 5th July

  23. Near Future • Given these problems we have started work on the later deliverables. • Major rewriting of the build procedure will help with inclusion in GENSER. • We have had significant interactions with the LCG GENSER project • number of talks at the monthly GENSER meetings • regular phone/VRVS meetings with members of the GENSER project. GridPP, Durham 5th July

  24. Near Future • Changes have been made to allow compilation on CERN systems, workarounds for compiler bugs. • Next major project is the study the inclusion in the GENSER libraries and the SCRAM installation system. GridPP, Durham 5th July

  25. Interfacing with EvtGen • The most important physics deliverable is the interface with EvtGen. • Work has started on this using unfunded effort. • Given the sophisticated nature of both decay simulations this is a complicated task. • Not just a matter of passing momenta and particle ids. GridPP, Durham 5th July

  26. Interfacing with EvtGen GridPP, Durham 5th July

  27. Interfacing with EvtGen • Angle of outgoing B meson with respect to the beam in needs careful handling to get the correct correlations GridPP, Durham 5th July

  28. Conclusions • We are making progress. • It has been a lot slower and more painful than I had hoped. • I think we will succeed with the majority of the deliverables. • However many will be late due to factors beyond our control. GridPP, Durham 5th July

More Related