• Presentation on the current status by Dorothea vom Bruch
• In Alice: Kalman filter running on GPU in single precision, including extrapolator, but they have a solenoidal magnetic field -> much easier than LHCb's, but there are some special cases to avoid problems in single precision -> what can we learn from them?
• Many people who are interested in contributing
  • Andrey's Yandex group: would like to work on muon ID, more people in September
  • Mike Sokoloff (Cincinnati): ML algorithms for HLT1
  • Arantza's group (Valencia): contribution to be defined
  • Brij (Valencia): improve downstream tracking performance, add to HLT1 sequence
  • Diego (Santiago) interested in improvements in downstream tracking
  • Florian (LPNHE, Paris): PV finding
  • Vava (LPNHE, Paris): Forward tracking on x86 in cuda_hlt1
  • Jacopo (CERN): forward tracking on GPU
  • Xavier's group (Barcelona): interested in ML techniques
  • Placido (CERN): VeloUT performance improvement on GPU, interested in memory management
  • Roel (Nikhev): framework integration: conditions, magnetic field

-- DorotheaVomBruch - 2018-07-23