Repository navigation
To test
andy edited this page Oct 27, 2026
·
97 revisions
Last updated: 2026-10-27 22:55 CET. Hand checks waiting for a person at the screen. An item is added when a fix is merged and removed when it has been checked; anything that fails goes back to its issue. "In build" is the first commit that contains the fix: the package you test must be that build or later.
How to help: anyone can take an item. Install the build named (packages from the latest
CI run on main, or build as in GETTING_STARTED.md), follow
"How to test", and comment on the linked issue: what you did, what you saw, operating system and version, and a
screenshot if something looks wrong.
| Check | How to test | What to look for | Issue | In build |
|---|---|---|---|---|
| GROMACS launch messages | GROMACS Dynamics: Dynamics tab, constant pressure on, pressure relaxation time 0.5 ps; Run settings, barostat interval 100; time step 2 fs (the default). Launch | The launch check stops on grompp's warning; the message says 'Dynamics tab, pressure relaxation time (tau-p)' and 'Run settings, barostat interval (nstpcouple)' instead of bare .mdp keys, and suggests a larger relaxation time | #527 | 169c1982 |
| Allocation profiles | Register Machines, KebnekaiseX, Machine tab: Allocation profiles, Add (e.g. Project with the hpc2n2026-351 account), Edit, Delete; then the Launcher | The list shows Default first (editable, not deletable); the Launcher shows a Profile choice only with more than one profile, and its fields fill from the chosen one | #541 | db3fd002 |
| Allocation left | Organizer: select a running Kebnekaise calculation (with the ssh connection open) | Under Comp. Date: Allocation: and Left: … (checked hh:mm); also for jobs launched before today (account from the profile) | #613 | db3fd002 |
| Ar drawn, water whole | Delete Examples/General_Chemistry, File > Open Example Calculations > General Chemistry, wait for the import to finish; open ar-gas, then water-spce | Argon atoms visible (pale, white); water molecules whole, no bonds across the box | #560 | deb 128310c8 (10-27 13:35) |
| Queued job: updated | Launch a job that waits in the Kebnekaise queue (or Reconnect a queued one with this build); select it in the Organizer | The Progress box shows 'Queued since …', the job id and 'updated n s ago', which goes back to a few seconds about once a minute | #606 | 1678bf87 |
| Hessian assembly line | ORCA Opt+Freq with analytic frequencies; watch the Progress box while it assembles the Hessian | 'analytic Hessian, n of n perturbations · assembling the Hessian', nothing after it | — | 7b90f6b8 |
| ORCA opt not converged | ORCA Opt with a small cycle limit (e.g. %geom MaxIter 3 end in the input), Launch | Ends Unsuccessful; summary reason quotes 'The optimization did not converge' | #614 | d1c18d47 |
| Remote Browse | Register Machines, a remote machine (e.g. NerdEgg), Codes tab: Browse next to a program path | A chooser lists that machine's folders (Up, path box, double-click a folder to enter, a file to choose); the chosen path lands in the Program box. On localhost Browse still opens the normal file dialog | — | c359e84a |
| Return in message boxes | File > Exit (Quit ECCE dialog), press Return | Return quits (the highlighted button), it does not cancel | #634 | 4166259a |
| Launch Selected, new machine | With the Organizer open, register a new machine in Register Machines; select calculations, File > Launch Selected, Launch on the new machine | No 'not registered here'; the calculations are planned for it | #588 | 28a01239 |
| CP2K 2023 optimisation progress | Select a running (or Reconnect a) CP2K 2023 geometry optimisation, e.g. nbomat/03_SOMS_…_opt | The Progress box shows the optimisation step, energy, max/RMS gradient and step against their limits, with the force plot | — | ed799a93 |
| Request lines per queue | Register Machines, Kebnekaise, Queues tab: put #SBATCH --constraint=zen4 in Request lines, Update Queue, Save; Launcher on that queue: job script preview; launch a short job |
The line is in the preview and in the submitted script; the job starts on a zen4 node. A Slurm machine without the memory option (KebnekaiseZ4) shows Memory in the Launcher | #624 | 8024cb5a |
| Summary after the allocation answer | Viewer of a Kebnekaise calculation (Allocation shown), wait a few seconds for 'Left' | Labels stay on the left, values on the right; the long Run Directory wraps, the panel keeps its width | — | f2add281 |
| scancel on the cluster | Launch to Kebnekaise, then scancel the job on the cluster while it is pending | Within a minute or so the calculation shows Killed, reason 'cancelled on …'; Reconnect not needed | #636 | ef9daa18 |
| Launch on another machine | Select several calculations set up for Kebnekaise, Run Mgmt > Launch Selected, Launch on: radium | Changed column lists queue/cores/time changes; after Launch the calculations show radium as their machine | — | bfaa93cb |
| Frequency counts | Any ORCA, Gaussian 16, NWChem or MOPAC Opt+Freq that finished | Progress box / summary: 'n real, m imaginary'; next to converged, the imaginary count, red when not 0 for a minimum | #611 | 2fe4ca58 |
| macOS GROMACS examples | Mac with a dmg from 85c44ab5 or later: delete Examples/General_Chemistry, reopen the set, open ar-liquid | Atoms drawn, Geometry Trace and MD analysis present | #630 | 85c44ab5 |
| Density Profile pick on a cluster | 01_Nb10O28Li_5m_s1_orca_optfreq: Density Profile, click one atom, then a second | No freeze: the first click sets the line's first atom, the second draws the line | #629 | deb 128310c8 (10-27 13:35) |
| Density Profile pick on an atom | Any single atom or diatomic (e.g. General Chemistry atoms): Density Profile, click an atom | Opens the profile around that atom, as before | #629 | deb 128310c8 (10-27 13:35) |
| Check | How to test | What to look for | Issue | In build |
|---|
| Check | How to test | What to look for | Issue | In build |
|---|