Hi Rouven,

We still use MadGraph4 for a couple of things, and it's written in fortran, mainly using the f77/g77 compiler by default. However, we now only have gfortran at JLab after the centos7 upgrade, so the compiler had to be changed everywhere to get our existing generators to work.

Changing all the instances of g77 to gfortran in a MG4 generator, and trying to get everything to compile, I get a lot of "Warning: Character length of actual argument shorter than of dummy argument", and is probably related the following problem, when generating events:

Working on subprocess:
    P1_e-n+_f+f-e-n+    cuts.f:21.28:

      pass_point = passcuts(p)
                            1
Warning: Rank mismatch in argument 'p' at (1) (rank-2 and scalar)
genps.f:54.53:

      call gen_mom(iconfig,mincfig,maxcfig,invar,wgt,x,p)
                                                     1
Warning: Actual argument contains too few elements for dummy argument 'x'

So there is a tensor/scalar rank mismatch that creates a zero cross section, and I don't know what's causing it. Everything builds, except Subprocesses, where genps.f is located:
"No rule to make target `nexternal.inc', needed by `genps.inc'"

So variables in there may not be getting set.

This generator is in
/scratch/hps/byale/MG/ap/MG_mini
on ifarm1402, if you'd like to try running it.
Do you know what could be causing this?

-Bradley