HSF Weekly Meeting #135, 4 May, 2018
Present/Contributors: Graeme Stewart, Benedikt Hegner, Pere Mato, Dario Menasce, Scott Snyder, Michel Jouvan, Guilherme Amadio, Ben Morgan, Pete Elmer, Simone Campana, Sudhir Malik
News, general matters
- Two abstracts were submitted to the RSE 2018
Conference
- One general one on HSF as an organisation
- One specific one from the Packaging WG
- Should hear the outcome quite soon
- IEEE Nuclear Science
Symposium in Sydney,
Australia, 10-17 November 2018
- Abstract submission open until 9 May
- IEEE eScience
Conference in
Amsterdam 29 October - 1 November, 2018
- Abstract submission until 25 May
- There is a particularly relevant track on Exascale Computing
for High Energy Physics
HSF/WLCG Workshop Follow-up
- Practical follow ups at the HSF level
- Inventory of community activities
- Graeme will work on this real soon now!
- More variation in themes in these meetings (more topical)
- See news on packaging below
- Collaboration guide for projects
- Draft by Benedikt. Iterating with Graeme
- Gordon Watts has proposed UW Seattle for the next workshop
- Next general meeting would probably be in ~one year from the
HSF side
- Need to discuss with WLCG colleagues
PyHEP Workshop
- Attendance still increasing … and we welcome more attendees, of
course!
- Status of AGENDA, largely finalised:
Saturday, 7 July
09:00 -\ 09:10 Welcome and workshop overview
09:10 -\ 10:30 Historical perspective / overview
- The slow adoption of Python in HEP
- The Python scientific software ecosystem - past, present and future
- Open discussion, guided by some survey input
11:00 -\ 12:30 HEP python software ecosystem
- The Python ecosystem in HEP data analysis
- Python packages for Machine Learning (TBC)
- GPU-oriented Python analysis package example (TBC)
12:30 -\ 14:00 Lunch
14:00 -\ 15:00 HEP python software ecosystem
- Python and Deep Learning for Neutrino Experiments (TBC)
- The Scikit-HEP project
15:00 -\ 16:00 Analysis & HEP frameworks
- The NEXT experiment analysis and data flow
- The Belle-II analysis framework
16:00 -\ 16:30 Coffee/tea break
16:30 -\ 17:30 Distribution and installation
- LCG releases: a complete, mult-ilanguage analysis ecosystem
- Distributing Python for the HEP environment (TBC)
17:30 -\ 18:30 PyROOT and Python bindings
- Tools to bind to Python
- ROOT’s C++ Python bindings
Sunday, 8 July
09:00 -\ 10:40 Analysis & HEP frameworks
- KEYNOTE PRESENTATION on JupyterLab
- Python tools for simulating beam dynamics
- Python for "core software" in ATLAS
- Python for “core software” in CMS
10:40 -\ 11:00 Coffee/tea break
11:00 -\ 12:00 Python 2 versus 3
- Python 2 versus 3
- Discussion on community plans for the future
12:00 -\ 13:00 Open discussion on education and training
13:00 -\ 14:00 Lunch
Activity updates
Licensing
- Meeting organised for next
week, 9 May at
17h CERN time
- ATLAS and LHCb will give an update on their © and license status
- Belle II will report on their deliberations and open questions
- FastJet told us they will relicense to GPLv2+
- This is an improvement over GPLv2 only
- But still copy-left, so substantial impact on users’ software
- AIDA2020 WP3 (Software) meeting discussed software licenses, and a
move from GPLv3 to LGPLv3, summarised in the closing
talk
from Frank Gaede
- “in principle should be able to re-license relevant packages;
need agreement of all copyright holders…”
- Doesn’t seem we have a licensing list, should we?
- Pere and Michel support a more focused list for discussions; we
will set this up for next week
Training
- We had a kick-off meeting last Monday with the goal of trying to
create a “federation” of Schools of Computing. 12 people
attended, some of them representing existing schools some other
just representing users. We introduced ourselves and our roles.
-
URL for minutes of Training and Career Working group
meeting here
Dario illustrated the goal of this initiative and its potential
benefits (improved efficiency and cost-effectiveness of having a
coordination among existing schools). No significant objections were
made, so we consider the idea worthwhile to pursue and start building
around it.
First step consists in putting together a mini DB capturing the
current situation (existing initiatives, goals, existing material,
future programs,...). This is being done on
https://docs.google.com/spreadsheets/d/1J5DVs0f1aUpsfp5hewGos43NCr7jr3unWMH8kivkKsc/edit#gid=0
(please anyone, contribute with your piece of additional information
if you have any).
This document should later evolve in a dedicated web page (under the
HSF site tree) to show training schedules, public
material-presentations/good software related examples/exercises and
best practices/reference material.
Identified a three tier structure to work on and later develop:
- Lower - examples - Linux/C++/Python
- Medium – ROOT/Geant4
- High – Machine Learning/ GPUs/ Developer level
Foreseen a regular calendar of future meetings (still to setup and
decide)
- Pete - many basic things are part of Software Carpentry, so we
should try to re-use that. Will we need to augment it as well?
E.g., data analysis for ROOT. Get a consensus around this basic
material.
- Dario - this wasn’t really discussed, agree with the thrust of
the idea. Input from Universities is also important.
Packaging
- Meeting May 2
covered micro-architecture builds and progress in Spack from FNAL,
minutes
available.
- Next meeting will be May
16, should
review progress on Test Driving.
- Updates on the packaging repo in GitHub to help people test
this.
- Would be good to update the note on micro-architectures.
Frameworks
- Frameworks Birds of a Feather session at CHEP
GSoC
- Report in the next meeting from organisers.
CWP
-
General Matters and Roadmap
- Agreed with editors that CWP Roadmap can be submitted to CSBS,
Graeme and Michel will take care of preparing the submission
in the proper LaTeX format (quite some work may be needed!).
-
Publication strategy for Individual WG Papers
-
Machine Learning
-
Software Trigger and Event Reconstruction
- Just needs PR to hsf-documents.
-
Data Organisation, Management and Access
- Text finalized, submission about to be done by Bo.
-
Data and Software Preservation
-
Visualization
-
Event/Data Processing Frameworks
- Still needs one more iteration.
-
Careers, Staffing and Training
- Would like to verify with the community whether to include
in the document the idea of a federation of existing (as
well as new one) schools of computing.
-
Facilities and Distributed Computing
- WLCG strategy document is now released. This captures a lot
of information. Ian B will need to decide how to proceed.
-
Conditions Access
AOB
- No meeting next week (Ascension holiday in a lot of Europe).
Reconvene on 17 May.