- Compact style
- Indico style
- Indico style - inline minutes
- Indico style - numbered
- Indico style - numbered + minutes
- Indico Weeks View
# Tue 24.09.2024 CMS/PhD meeting
https://indico.cern.ch/event/1373477
Present: RS, ZW, AV (notes), DM, SR, OM
Excused: Sapta, Jin
## Reweighting (Zenny)
ZW presents some slides about the teawrex reweighting.
And some interactive demo.
All: very good work!
Discussion about how to go forward.
SR: we have a release LO upcoming first, should discuss with OM
RS: we should keep the same interface as the rest of madgraph
RS: we can validate this earlier than the normal CMS release cycle if needed
RS: can we use this for pdf reweighting?
ZW: should look into this with OM
RS: will discuss with gen level2 conveners
RS: maybe we could invite him here to this meeting, you repeat the presentation
SR: please RS take care of inviting the right people
SR: next tuesday, so one week from now
AV: looking at technicalities, we should discuss in the dev meeting
See https://github.com/madgraph5/madgraph4gpu/compare/master...zeniheisser:madgraph4gpu:rexDev
AV: 1 about packaging, do we want one package or two? can also be one... I am just working on this with OM now
AV: 2 if one package, better in any case to have mnodularity: separate the output.py and model_handling.py parts for cudacpp and trex into separate files
AV: 3 another suggestion for modularity (but not strictly needed, if this is one package): have cudacpp libraries built by the cudacpp (eg SA) makefile and then trex use those libraries
AV: 4 very nice that the rest of cudacpp looks almost untouched, except for debugging... anyway should ensure that adding trex does not change the behaviour of the rest of cudacpp
[AV after the meeting: 5 it would be nice to have tests of trex in the CI]