Inder Monga
(ESnet),
Kaushik De
(UTA),
Shawn McKee
(UMich),
Torre Wenaus
(BNL)
26/09/2019, 14:00
Includes ESnet/ATLAS/CMS network studies, iDDS and other data lake initiatives, potentially remote data access from Tape Carousel.
Glen Lockwood
(NERSC),
Peter van Gemmeren
(ANL)
26/09/2019, 15:30
A discussion on I/O challenges for HPC production jobs. Includes Event I/O of course (TMPIRoot, HDF5, object stores like HEPnOS, etc) but also DB access for reconstruction jobs, lustre and small files. containers and job initialization.
Brian Friesen
(NERSC),
Charles Leggett
(LBNL)
26/09/2019, 16:30
What is the goal? Reduce latency when running trigger algs, increase throughput on HPCs
What are the challenges? Insufficient parallelism, code size, ...
Examples: track seeding, Fast Chain, Geant X
CCE Portable Parallelization Strategies
ML algorithms (Jet Classification, Tracking,...)
...
Doug Benjamin
(ANL),
Miha Muskinja,
Rollin Thomas
(NERSC)
26/09/2019, 17:20
event service, raythena, yoda, pilot 2, jumbo jobs and their interactions
