Merge lp:~maddevelopers/mg5amcnlo/MadSpin_correctbug into lp:mg5amcnlo/lts
Proposed by
Pierre Artoisenet
Status: | Merged |
---|---|
Merged at revision: | 258 |
Proposed branch: | lp:~maddevelopers/mg5amcnlo/MadSpin_correctbug |
Merge into: | lp:mg5amcnlo/lts |
Diff against target: |
29 lines (+11/-2) 1 file modified
MadSpin/decay.py (+11/-2) |
To merge this branch: | bzr merge lp:~maddevelopers/mg5amcnlo/MadSpin_correctbug |
Related bugs: |
Reviewer | Review Type | Date Requested | Status |
---|---|---|---|
Olivier Mattelaer | Needs Fixing | ||
Review via email: mp+238311@code.launchpad.net |
Description of the change
Hi Olivier,
This is the correction to the bug
https:/
For some reason the status of the random generator was not properly
written down before each termination of the fortran code in the unweighting mode.
I have checked that the ranmar state is properly written down in this version.
Following Markus suggestion, I also set the max number of calculators to a low number
and check the absence of spikes in the distributions (for pp > tt~ j, all decay channels)
Cheers,
Pierre
To post a comment you must log in.
Hi Pierre,
I agree that your fix is working, but I think that we need to slightly secure it for the random seed problem in the future.
So here is my suggestions: matrix_ element" stdout. readline( ))
nb_ output = abs(info)+1
nb_ output = 1
1) add an check that the routine "calculate_
is only use in mode "decay"
i.e.
replace the lines:
if mode == 'prod':
info = int(external.
else:
info = 1
by
if mode != 'prod':
raise Exception, "Mode not supported anymore in this function (not save for seed)"
info = 1
nb_output = 1
2) having a nice treatment of that part as well.
This is of course very small but will make me sleep better ;-)
Cheers,
Olivier