fix patch to be compatible with mg5 branch gpucpp_goodhel

[oliviermattelaer]

This is a simple patch to make the plugin work for the gpucpp_goodhel branch
(removing the double list for the helicity initialization and allowing limhel to be control from the run_card.)

The fact that limhel is now a user parameter makes it important to be correctly handle by the cudacpp interface as well which is not done in this PR.

[oliviermattelaer]

@valassi

The PR obviously now crashes since limhel !=0 since now that parameter is part of the run_card setup.
I can do a quick/temporary fix which is setting that parameter in the run_card by default when doing output in the simd/cuda mode to have this PR going trough quickly.
But I guess that we need to

1. change the test to enforce limhel=0
2. change cudacpp to support correctly limhel (even if the algorithm is not 100% identical to the fortran one, like when this was done via a running sum rather than comparing one helicity to the average of the helicity for that event (like done in fortran).

Do you want to do 1 or 2 directly? or should I implement the different default value?
(which is super easy to do)

[oliviermattelaer]

I had to remove a useless call to reset_cumulative_call in the cudacpp since now I discard those in the fortran.
And now all tests are passing.
Just we can either review/merge this or first fix the test and put back in cudacpp some support of limhel.

(maybe we can skype on this tommorow --if you are available--)

[oliviermattelaer]

Hi,

Just to comment that the check that MC over helicity was not used for the cudacpp plugin was already in place.

So do not need to do anything here.

Olivier

[valassi]

Hi @oliviermattelaer this is goo dto go for me, but first we need to mertge master into it and fix conflicts. This can be done by merging #986 into this. Can I go ahead?

[valassi]

This #955 'goodhel' as well as the additional #986 'additional goodhel' (which would be needed to fix the master conflicts) are included in the #992 which is about to be merged. I close this for simplicity.

In any case I have marked that #992 fixes also the issues #872 and #950 that are fixed in this #986 and the underlying #955

Closing this for simplicity. All this work will be merged via #992.