Muon Week 2017 Run3 - MRRTF
Laurent Aphecetche
19 May 2017
Reminder of the problem at hand
Going from
8 kHz PbPb @ 1 kHz w/ triggers
To
50 kHz PbPb @ 50 kHz triggerless
Lots of data
3 TB/s CRU
90 GB/s
Must fit in
Time budget : treat data fast
Storage budget : compress data
Alice (TPC) priority
Data compression
Muon-only needs are more modest
~ few GB/s
envision full online reco
O2 Data Flow
Frames everywhere
ALFA
Part of O2 is based on ALice FAir framework
- working unit is a device
- devices communicate through messages
- devices can be :
- single or multi-core
- on same machine or on different machines
- in different languages
General O2 news
Bye bye Computing Working Groups
Hello Work Packages
WPs Institutional responsabilities
WP descriptions and planning
- WP1 (data model)
- WP3 (common tools and software infrastructure)
- WP4 (O2 software framework)
- WP11 (ALFA)
- WP12 (detector simulation)
- WP13 (reconstruction and calibration)
Also relevant for some of us
- WP6 (detector readout)
- WP7 (data quality control)
- WP9 (event display)
- WP10 (CCDB)
- WP14 (analysis framework and facilities)
- WP15 (data management)
O2 links
-
- plenary 1/month
- technical 1/month (alternate)
- WPs meetings (1/1-2 weeks)
MRRTF
Muon Run3 Reconstruction Task Force
Current cast
- Diego 30%
- Gabriele 40%
- Laurent 60%
- Philippe 20%
- Hugo 30%
- Ivana 20%
- Bogdan 20%
- Sebastien 30%
- Sean 40%
...FTE numbers are just wishes for the moment...
1st priority
Reconstruction
2nd priority
Simulation
| What | Event | Start | End |
|---|---|---|---|
| MRRTF | init | Feb 15, 2014 | Apr 15, 2014 |
| MRRTF | Kickoff meeting | Oct 14, 2014 | Oct 16, 2014 |
| MRRTF | 1st hackathon | Mar 8, 2016 | Mar 11, 2016 |
| MRRTF | 2nd hackathon | Oct 18, 2016 | Oct 21, 2016 |
| MRRTF | 3rd (virtual only) | Feb 6, 2017 | Feb 9, 2017 |
| MRRTF | 4th hackathon | May 9, 2017 | May 12, 2017 |
| O2 | TDR | Feb 1, 2014 | Apr 1, 2015 |
| O2 | Work packages | Mar 1, 2017 | Dec 15, 2021 |
| O2 | Simu challenge | Jan 1, 2018 | Dec 31, 2018 |
| O2 | 10% data challenge | Jan 1, 2019 | Dec 31, 2019 |
| O2 | Rec+sim (ITS,TPC) | Feb 1, 2017 | Dec 31, 2017 |
| LHC | LS1 | Feb 16, 2013 | Dec 17, 2014 |
| LHC | LS2 | Dec 31, 2018 | Jan 1, 2021 |
| LHC | Run3 data taking | May 1, 2021 | Dec 15, 2021 |
2020
Main issue so far is ...
(wo)manpower
how could you possibly guess ?
Availability of manpower to be more precise
We do have volunteers
but some had little free time so far
Net result being we're (almost) not moving
Remedies ?
1 - Carve some time out for volunteers by finding other people to do part of their job
Example @ Subatech : trying to hire a post-doc focused on physics to free some time for permanent physicists
Remedies ?
2 - Hire some skilled people (PhD or otherwise) for a 2-3 years to do part of the job
Easier said than done ?
Some progress
Go FairMQ device
(March Alice Offline Week, Sebastien)
Demonstrates FairMQ-like device can be written in a different (i.e. not C++) language
WIP
alo
sampa go
mid occ
mrrtf "organization" on
host code before integration to AliceO2 repo
mrrtf on the
meant as an information tool for the team
JIRA "Agile" planning
or
BACKUP
1
Muon Week 2017 Run3 - MRRTF
Laurent Aphecetche
19 May 2017