different behaviour with nagfor and gfortran concerning phase space configuration

[sschmidt]

When having multiple integrate commands in one sin file, there is different behaviour: compiled with nagfor the phase space is regenerated before every integrate command -> each integration therefore gives slightly different results compiled with gfortran the 'existing phase space configuration' is reused -> each integration gives the same results

I'm not sure which behaviour is intended

[sschmidt]

simple example with part of the output

[kilian]

I guess that one run was with and the other without the -r option.

-r implies ?rebuild_phase_space=true initially. You can reset this anywhere in the script, or simply omit -r.

Closing the ticket.

[sschmidt]

I can't reproduce it now. But I also don't remember setting any -r flags...

So I just leave it closed...

[Juergen Reuter]

Milestone v2.0-rc2 deleted