BUG FIX in PR #47: arrays were uninitialised in RAII constructors

[valassi]

As described in #73 this is a spinoff of #72

[roiser]

The difference between the two lines should be that the new one initializes the array values all to 0. I just wonder where we depend on 0 inits in the code? Could you point me to the line?

But of course it's fine to have proper 0 initializations, therefore approve it.

[valassi]

The difference between the two lines should be that the new one initializes the array values all to 0. I just wonder where we depend on 0 inits in the code? Could you point me to the line?

But of course it's fine to have proper 0 initializations, therefore approve it.

Hi Stefan, this has showed up in my latest changes in #71. Amongst other things I reimplemented the helicity filtering in c++ in the same way as in cuda. I was initially getting a worse performance, because some bad helicities were not filtered out. This is from https://github.com/valassi/madgraph4gpu/blob/acc333ac82223842162dcc6e4bf39bb802eafd19/epoch1/cuda/ee_mumu/SubProcesses/P1_Sigma_sm_epem_mupmum/CPPProcess.cc#L724, the missing initialization being in https://github.com/valassi/madgraph4gpu/blob/acc333ac82223842162dcc6e4bf39bb802eafd19/epoch1/cuda/ee_mumu/SubProcesses/P1_Sigma_sm_epem_mupmum/check.cc#L213

Anyway, looks like you approved :-)

Thanks, I will merge
Andrea