#451 closed feature_request (fixed)
Introduce run IDs
Reported by: | kilian | Owned by: | kilian |
---|---|---|---|
Priority: | P3 | Milestone: | |
Component: | core | Version: | 2.0.6 |
Severity: | normal | Keywords: | |
Cc: |
Description
When scanning over parameters, integrating the same process repeatedly, earlier results are always overwritten. In a second run, Whizard should be able to shortcut this as usual. Instead, it has to redo everything.
(Unique) run IDs would eliminate this nuisance. They should be introduced behind the scenes, if possible.
Change History (4)
comment:1 Changed 13 years ago by
comment:2 Changed 13 years ago by
Milestone: | v2.0.7 → v2.0.8 |
---|
comment:3 Changed 13 years ago by
Resolution: | → fixed |
---|---|
Status: | new → closed |
Done in r3757. Documentation is in the updated manual.
(This revision fixes also a problem with OpenMP in test/unit matrix elements, which don't use OMega.)
Note: See
TracTickets for help on using
tickets.
I ran into this repeatedly, when I tried to scan over parameters, but tried to keep the individual integrations grids and event files. A dirty way out of this would be to use a different process name for each run, in that the grids could be reused, if the complete scan is run again. Unfortunately, assigning strings or integers as process names isn't working in SINDARIN (by design?). Alternatively, saving the grid's to a specific file, as in the "sample" case for events, could also do the job. Nonetheless, run ID's are presumably the best way out.