[Ended] Adaptyst CERN sprint #7

Sprint number: 7
Dates: 27/08/2026 - 09/09/2026 11/09/2026 (extended due to the vacation of the main R&D person in the week of 14/09)

Issues:

  • Move from patched “perf” to in-house equivalent: to do
  • Solve lock contention / too large communication overhead problem in Adaptyst API methods*: done
  • Fix nvgpu bugs leading to inaccurate results (like CUDA runtimes taking >100% of a region): done (new bug(s) discovered and fixed)
  • Fix group label formatting in canvas-timeline*: done

*AI assistance will be used in the context of this task.

Some issues may be hidden: if you have permissions (e.g. a CERN affiliation), you can view them by logging in to the issue tracker.


Glossary of possible issue states:

  • Backlog: not scheduled in a currently-ongoing sprint.
  • To do: scheduled in a currently-ongoing sprint, but no work has begun yet.
  • In progress: the work is being carried out at the moment.
  • On hold: the work is being carried out in general, but it is not scheduled in a currently-ongoing sprint.
  • Waiting for third party: waiting for an input from another person/group at CERN.
  • Waiting for external: waiting for an input from a non-CERN person (e.g. a collaborator or an external contributor like you). We need to stress here that voluntary contributors never have any commitments towards us and are perfectly allowed to leave at any time: this is absolutely fine for us, but we’ll appreciate letting us know in this case so that we can update our planning.
  • Done: completed.