comprehensive set of fixes for SUSY (complete) and EFT (partial) - keep susy_gg_tt, susy_gg_t1t1, heft_gg_h in repo and test them in CI

[valassi]

Hi @oliviermattelaer I have completed many more fixes for SUSY so I am including them here.

This PR includes also all the changes that are in PR #625 (first set of susy changes) and in PR #822 (fix for a helas bug affecting also non SUSY processes). I filed separate PRs to make it easier (maybe) to review.

But if you prefer, I can also just merge this straight away and close the other two.

Can you please review?

Notes

• This closes SUSY HRDCOD=0 builds in CUDA give MEs equal to 0 (warning in parameter access through getInstance in device code) #818 aand SUSY HRDCOD=1 builds fail because sin/cos/atan are not constexpr (only CUDA fails, while C++ succeeds) #627 (through the changes in PR fix susy gg_tt builds and add it to the repo #625)
• This closes Bug in GPUFOHelasCallWriter format_coupling (wrong params2order for SUSY processes) #821 (helas bug, through the changes in PR Fix bug in GPUFOHelasCallWriter format_coupling #822)
• It also closes Out-of-bounds array warning in SUSY gq_ttllq #823 (out of bound arrays bug in cIPC), not in any other PR
• And it especially closes all of the issues in Further issues in builds of other SUSY processes #820, not in any other PR

The status is: all SUSY processes that I tried are now correctly code generated, built and tested. Can you maybe also suggest new processes to try?

Thanks
Andrea

cc @roiser

[valassi]

Ok I have completed my tests.
@oliviermattelaer I think this is ready to be merged when it is reviewed.

Note, I have also implemented the "tmad" tests comparing fortran to cuda/cpp, an dthey fail for both susy_gg_tt #825 and for susy_gg_t1t1 #826. But I suggest meging this one already (which fixes many other things!) and then addressing those other issues later.

By the way, again, eventually I'd like to remove heft_gg_h by something more interesting, since heft seems fixed (unlike smeft).

[oliviermattelaer]

Sounds good some minor point to check -> let's say approve!!!!
Thank you

[valassi]

I will merge to ease the comparison to smeft and then heft.

Keep here the three comments for later:

• what about param.name == 'ZERO' ?
• mdl_g will be return true for mdl_g1 as expression?
• check the status of "FIXME! Here there should be different code generated depending on MGONGPUCPP_NBSMINDEPPARAM_GT_0"?

[valassi]

I will merge to ease the comparison to smeft and then heft.

Keep here the three comments for later:

* what about param.name == 'ZERO' ?

* mdl_g will be return true for mdl_g1 as expression?

* check the status of "FIXME! Here there should be different code generated depending on MGONGPUCPP_NBSMINDEPPARAM_GT_0"?

Ok I have added comments on all three points (previously resolved without answers). All done for me.