SamWiseMinutes3 6 14¶
Attendance: Nate, Andrew, Craig, Kanika, Adam, Gavin, Jan, Dominick, Eric, others…
-- Need a tag — soon?
-- Dominick: Run PChits to get files ready for calibration studies.
-- Adam: Modifications to metadata module need for CAF metadata
-- Adam and Dominick work together to use Adam's ArtSAMWrap modification.
Action items from last time:
Dominick: PC Hits:
--> It is running okay – hitting bad channels the way it is supposed to. Is the channel mask sane?
--> Need a tag and then we can submit jobs in large numbers.
Meet with some of the new people to get them started with daq2rawdigit:
- (Siva Prasad-Kasetti, Kuldeep Kaur-Mann, Pavan Poot Pandey and possibly Nitin Yadev.)
--> talked to Siva and Pavan on Monday. and we will continue working on getting them up and running over the week.
- metadata module update to get CAF going
--> The metadata for CAFs is almost done. I am implementing this by making the MetadataManager a singleton class. The caf maker module will be able to get an instance of the manager that the metadata module has been filling.
--> No exception if you run over a file that isn't in SAM.
Adam and Chris: follow up on multiple run sections in a single grid job:
--> See below. This is working for the case where a single grid job runs the "nova" command many times.
--> Multiple files per job, slow down at copy out. Single IFDH command only one lock..
--> Some changes in ArtSAM wrap accepted by Marc Mengel-- "-- multifile option"
--> Now, we have control over the loop over art jobs – so we can get the filenames before running the "nova" command.
--> Files still show up as being consumed.
Nate is busy for the next week.
--> Won't have time to work on a renaming script — merge with ArtSAMwrap
We can't run multiple files with the ArtSAMWrap version that is in our repository?
-- Adam has been working on a private one.
Dominick: We need to create a standard for our file names.
--> Uniqueness of file names is a requirement of SAM.
Jan: some troubles with IFDH tools:
-- Can't do IFDH_fetch from inside JobSub — IFDH_fetch is purely interactive.
-- IFDH_CP is what we should use from batch job.
-- IFDH_CP is giving some errors – running on FermiGrid using JobSub tools.
Production goals for this week:
-- MC generation requires new flux files — or to decide we won't use it this time.
--> Some problems with GENIE flux files and some of Robert's code.
--> Generation files can't use the new flux files.
--> Robert H. is helping with the solution. Raphael is also helping…
( fix is in nutools – could require a new tag)
--> Should talk about this at the convener's meeting today
-- Tag status:
--> Compilation and run time errors working with Jon over the last few days to get things working.
--> Thought we were almost ready – then everyone wants to add more things…
--> Almost ready. Make sure bad channel info is sane. Also, IFBEAMSPILL info throwing some exceptions.
--> Things are running and can be tested.
--> This tag would be for PChits (Gavin MC and Dominick for Data) - get new calibration constants and then will be ready for reconstruction of data and MC.
-- Time crunch to get data for the scintillator study.
-- What can be hope to accomplish by Collaboration meeting:
-- simulation group demands using gsimple for flux files.
-- PCHIT holding things up, but Chris can get calibration quickly.
-- RECO will be run on Fermigrid and offsite if possible.
(could be done in a week once we have calibration)
Week 1 -- light reco + calibration
Week 2 — full reco
Week 3 — PID/LEM
(New PID files – still under development)
Week 4 -- CAF
-- Not realistic to get everything in the CAF files by collaboration meeting.
-- More important to make sure the system works properly.
-- Collaboration meeting – demonstrate that we can work through all of the chains.
-- FD cosmics could be through all steps in two weeks. (10,000 FD files declared to SAM)
( generation took one day – FTS took a little while)
( Ruth will want a dedicate hadron physics set of files that will require a lot)
-- Eric still waiting for OSG group.