The Corsika database used to query primary cosmic showers for our cosmic simulation doesn't have enough statistics when running with only proton primaries. The result is that in many cases we end up sampling the same cosmic primary (the start position on the cosmic generation surface is random, but not its 4-momentum).
Also, a single job on the grid picks only one Corsika database file and samples only from that file: we should also make sure that one file has enough entries for the events requested in a single job.
We need to generate more Corsika showers from protons.
The Corsika database used to query primary cosmic showers for our cosmic simulation doesn't have enough statistics when running with only proton primaries. The result is that in many cases we end up sampling the same cosmic primary (the start position on the cosmic generation surface is random, but not its 4-momentum).
Also, a single job on the grid picks only one Corsika database file and samples only from that file: we should also make sure that one file has enough entries for the events requested in a single job.
We need to generate more Corsika showers from protons.