The commands to use are essentially: ./bin/mg5_aMC import model loop_sm generate p p > t t~ b b~ [QCD] output TTBB_LesHouches Then in the TTBB_LesHouches/Cards/param_card.dat change 1. the top quark mass from 173 GeV to 172.5 GeV (element 6 in "Block mass") 2. the bottom quark mass from 4.7 GeV to 4.75 GeV (element 5 in "Block mass") In the TTBB_LesHouches/Cards/run_card.dat change 1. The PDF (“pdlabel”) from ‘nn23nlo' to ‘lhapdf' 2. The PDF ID (“lhaid”) from ‘244600’ to ‘260400’ (please double check this number with the LHAPDF website!) 3. Set the “parton_shower” from ‘HERWIG6’ to ‘PYTHIA8’ In the TTBB_LesHouches/SubProcesses/setscales.f you’ll have to change the default renormalisation scale to _geom. This can be done by replacing (line 223): parameter (imurtype=1) by parameter (imurtype=4) and adding (at around line 330) elseif(imurtype.eq.4)then tmp=1d0 do i=nincoming+1,nexternal-1 tmp=tmp*dsqrt(max(0d0,(pp(0,i)+pp(3,i))*(pp(0,i)-pp(3,i)))) enddo tmp=tmp**0.25d0 temp_scale_id=‘ET_geom of tops and bottoms' so that you have the structure if (imurtype.eq.1) then … elseif(imurtype.eq.2) then … elseif(imurtype.eq.3) then … elseif(imurtype.eq.4) then tmp=1d0 do i=nincoming+1,nexternal-1 tmp=tmp*dsqrt((pp(0,i)+pp(3,i))*(pp(0,i)-pp(3,i))) enddo tmp=tmp**0.25d0 temp_scale_id=‘_geom (special for ttbb)' else … endif That should basically be all to get the results with the default resummation scale. Using H_T/2 as resummation scale is not really recommended, and I wouldn’t know how the code would behave. To vary the resummation scale, you’ll have to change the TTBB_LesHouches/SubProcesses/mcatnlo_madfks.inc file. In particular the c Define lower and upper veto range (see MC subtraction terms) double precision frac_low,frac_upp parameter (frac_low=0.1d0) parameter (frac_upp=1.0d0) should be moved up and down (together) by a factor 2 to do the variation. Make sure to delete the object files (“ rm TTBB_LesHouches/SubProcesses/P*/*.o ") before generating the events for this variation, because I’m not sure this is done automatically when changing the madfks_mcatnlo.inc include file. Note that you’ll get the renormalisation and factorisation scale variation as extra weights in the Les Houches event file, so there is no need for separate running to get those. The top quarks can be decayed using MadSpin, selecting the final state you are interested in (by default, MadSpin includes all decays for the top quarks, with a diagonal CKM matrix). Remember to use the following settings when showering with pythia8: http://amcatnlo.web.cern.ch/amcatnlo/list_detailed2.htm#showersettings . These are automatically used when using the MG5_aMC build-in interface to pythia8. You might want to have a look at this https://answers.launchpad.net/mg5amcnlo/+faq/2613 to help in using the large number of machines effectively, although, it might not really be necessary for this process as it should not be extremely slow.