Changes between Version 48 and Version 49 of Reweight


Ignore:
Timestamp:
Apr 5, 2016, 1:04:17 PM (8 years ago)
Author:
Olivier Mattelaer
Comment:

--

Legend:

Unmodified
Added
Removed
Modified

  • Reweight

    v48 v49  
    66 The method works only if both the original and the new hypothesis give non-negligible contributions to the same parts of the phase-space.
    77 
    8  We support three types of reweightings, one for Leading Order (LO) samples, and two for Next-to-Leading Order (NLO) samples (dubbed Kamikaze Reweighting and NLO Reweighting).
     8 We support four types of reweightings, one for Leading Order (LO) samples, and three for Next-to-Leading Order (NLO) samples. One of them is NLO accurate (dubbed NLO re-weighting) and two are approximate using LO method (LO/Kamikaze Reweighting and Loop Improved)
    99
    1010 '''Leading Order'''[[BR]][[BR]]
    … …  
    6565 with the final weight computed by recombining these weights according to the prescription given in the NLO descriptions.
    6666
    67  This method is mainly intended for case where the virtual can not be computed as in the case of loop-induced production. As the LO/kamikaze re-weighting it is not NLO accurate. One can argue that it should be more accurate than the kamikaze method but this statement was never proven anywhere so far.
     67 This method is mainly intended for case where the virtual can not be computed as in the case where the Born of the new method is a loop-induced production (justifying the name). As the LO/kamikaze re-weighting it is not NLO accurate. One can argue that it should be more accurate than the kamikaze method but this statement was never proven anywhere so far.
    6868
    6969 '''This method will be released in a future version of MadGraph5_aMC@NLO''' and can currently be provided on request. Since it is based on a dedicated decomposition, the NLO sample must be generated in a specific way for the Les Houches event file to contain the necessary information (see below).
    … …  
    8181 This module is built-in in MadGraph5_aMC@NLO.2.3.2 and later.
    8282 It relies on '''f2py'''; the easiest way to install f2py is to install numpy (if not already done).
     83 For NLO and Loop-improved re-weighting it also relies on lhapdf.
    8384 
    8485== Running the code
    … …  
    137138}}}
    138139
    139 If you want to perform the NLO-accurate reweighting, you need to keep the parameter "keep_rwgt_info" to True.
    140 (This is set automatically to True).  With this option set equal to False (the default) the kamikaze reweighting will be performed.
     140If you want to perform the NLO-accurate reweighting, you might want to have the parameter "store_rwgt_info" of the run_card on True.
     141Note that this parameter will be automatically switched to True if you request at the time of the generation a NLO accurate re-weighting (or the loop-improved one).
     142You are force to set this parameter manually if you want to perform such re-weighting independently of the event generation (otherwise only the kamikaze reweighting will be available).
    141143
    142144
    … …  
    173175   5. '''change helicity <True|False>''': performs reweighting for the given helicity (True --default--) or carry out the sum over helicity (False).
    174176   6. '''change rwgt_dir <PATH>''': changes directory where the computation is performed. This can be used to avoid to recreate/recompile the fortran executable if pointing to a previously existing directory.
    175    7. '''change mode LO''': For NLO samples, this flag forces the code to perform the kamikaze reweighting (available in 2.4.0)
     177   7. '''change mode <LO|NLO|NLO_tree|LO+NLO>''': For LO sample, this line is always ignored. For NLO samples this select the reweighting mode "LO" for the kamikaze, "NLO" for the NLO accurate method, "NLO_tree" for the loop-improved and "LO+NLO" to run simultaneously the kamikaze and NLO method.
    176178
    177179 2. '''Benchmark definition''':[[BR]]
    … …  
    205207<initrwgt>
    206208<weightgroup type='mg_reweighting'>
    207 <weight id='mg_reweight_1'>set param_card dim6 1 100.0
     209<weight id='operator_1'>set param_card dim6 1 100.0
    208210</weight>
    209 <weight id='mg_reweight_2'>set param_card dim6 2 100.0
     211<weight id='mg_reweight_1'>set param_card dim6 2 100.0
    210212</weight>
    211 <weight id='mg_reweight_3'>set param_card dim6 3 100.0
     213<weight id='mg_reweight_2'>set param_card dim6 3 100.0
    212214</weight>
    213215</weightgroup>
    … …  
    227229       -4  1    4    4    0  502 +3.1115612e+01 -6.9750557e+00 +3.3618969e+02 3.37698598e+02 0.00000000e+00 0.0000e+00 1.0000e+00
    228230<rwgt>
    229 <wgt id='mg_reweight_1'> 4.55278761371e-06 </wgt>
    230 <wgt id='mg_reweight_2'> 2.65941887458e-06 </wgt>
    231 <wgt id='mg_reweight_3'> 8.68203803896e-06 </wgt>
     231<wgt id='operator_1'> 4.55278761371e-06 </wgt>
     232<wgt id='mg_reweight_1'> 2.65941887458e-06 </wgt>
     233<wgt id='mg_reweight_2'> 8.68203803896e-06 </wgt>
    232234</rwgt>
    233235</event>
    … …  
    235237 The above stems from a reweight_card that reads as follows:
    236238 {{{
    237 launch
     239launch --rwgt_name=operator_1
    238240   set Dim6 1 100
    239241   set Dim6 2 0
    … …  
    259261INFO: Original cross-section: 0.80086112072 +- 0.0025669959099 pb
    260262INFO: Computed cross-section:
    261 INFO: 119 : 5.0238030968
    262 INFO: 120 : 4.46724081967
    263 INFO: 121 : 0.790019392142
     263INFO: operator_1 : 5.0238030968
     264INFO: mg_reweight_1 : 4.46724081967
     265INFO: mg_reweight_2 : 0.790019392142
    264266}}}
    265267