LEF reading: Scale and alpha-QCD input

[kilian]

Apparently not yet functional. To be checked.

[kilian]

There's also the ?update_scale and ?update_alpha_s parameters which are currently disabled.

[Juergen Reuter]

?update_alpha_s and ?update_scale have benn used in 2.1 to change the alpha_s and scales for rescanning events. In 2.2 this seems to be done always if according expressions are specified. This seems more uer-friendly. So if really not needed these variables should be eliminated.

[kilian]

Replying to jr_reuter:

?update_alpha_s and ?update_scale have benn used in 2.1 to change the alpha_s and scales for rescanning events. In 2.2 this seems to be done always if according expressions are specified. This seems more uer-friendly. So if really not needed these variables should be eliminated.

Good, so remove those. I was not sure.

[Juergen Reuter]

Hm, I'm also not totally sure. Could someone double/triple-check? If confirmed, also ?update_parameters is irrelevant. What about the reading from LHE files?

[Juergen Reuter]

Yes, it does: update_sqme triggers evaluate_trace which contains via term_instance_evaluate_interaction the subroutine process_compute_amplitude: this resets alpha_s. So check. Directly before this routine call in process_instance_recover there is the call of term%instance%evaluate_expressions which expands and sets the scale expressions. SO they are modified if changed. This is called as event_recalculate (inherited to entry%recalculate in simulation%recalculate which is called both in cases of rescanning and of an alt_entry.

[Juergen Reuter]

I'm confused: is there something to be done now?

[Juergen Reuter]

Postponed to 2.2.1. Also partial overlop with #530 and #570.

[Juergen Reuter]

Actually, in the subroutine hepeup_to_event there is this code (commented out):

!!! Not implemented yet:
!     if (scale > 0)  call event%set_scales (scale)
!     if (alpha_qcd > 0)  call event%set_alpha_qcd (alpha_qcd)

Related to this?

[kilian]

While #701 is pending, I'll try to resolve this ticket. Plan: Parameters ?use_alphas_from_file and ?use_scale_from_file which tell WHIZARD to read the values in the file and use them in the calculation instead of the current environment setting. By default, the entries in the file are ignored. All of this applies only if ?update_sqme is true, and only if the file format supports those entries.

[kilian]

Enabled ?use_alpha_qcd_from_file for LHEF, in r6574.

To be done: enable for HEPMC and LCIO, and analogously for ?use_scale_from_file.

[kilian]

With r6578, the feature is implemented for LHEF, HepMC, and LCIO.

I'd close the ticket, except for

• how to name the new user-level parameters? I never know whether to write alpha_qcd or alpha_s (should probably stick to the latter).
• I believe that the LCIO code is correct (exactly analogous to HepMC) but couldn't test it, also didn't write unit tests - my runtime linker doesn't want to link the LCIO shared lib.

[Juergen Reuter]

Before you close the ticket, please do not forget to add the options to the manual, the main text and the SINDARIN reference list!

[kilian]

Replying to jr_reuter:

Before you close the ticket, please do not forget to add the options to the manual, the main text and the SINDARIN reference list!

Done in r6579. Also, reverted the parameter name to ?use_alpha_s_from_file, consistent with other parameters. I skip the missing LCIO test cases and close this ticket.