Peter Willendrup

Results 264 comments of Peter Willendrup

1st `GROUP` comp, raise a warning / recommendation: ``` INFO: No output directory specified (--dir) INFO: Using directory: "BAMBUS_20240527_145301" INFO: Regenerating c-file: BAMBUS.c Info: 'Al_window' is a contributed component. Warning:...

Seems to be fixed via https://github.com/McStasMcXtrace/McCode/pull/1813

(supplemented by https://github.com/McStasMcXtrace/McCode/pull/1814)

@Lomholy thanks for reporting/documenting this. Once you have fixed the bugs and decided on the best "test scenarios", I can recommend including `%Example` lines in the instrument headers (i.e. in...

True, this is built-in when using Union. It is on the other hand also on my list of possible improvements of the stand-alone PowderN that it could get this feature...

@mads-bertelsen thanks for noticing and reporting! For the sake of reproducibility, could you include your instrument files and settings used to generate the outputs? ( Also, once you get to...

@mads-bertelsen good point, this makes sense. I think the _better_ solution would in fact be that the initial tau_list structure became a component-induced `USERVAR`. This would of course result in...

Indeed in a CPU-only setting / memory-wise there is no difference between a `DECLARE` and `USERVARS` variable. But using `USERVARS` we outline the importance of keeping neutron-derived / calculated quantities...

Not a fully thought-through solution, but I guess something like this could potentially work - also on GPU - with relevant initialisation of numrefs and a `malloc()` in `TRACE` ```c...