Whizard slow down

Running Whizard for large samples of 10M events, we observe a slowdown of the output and a practical stop at 9M events
Example input:
[Muon.sin.txt](https://github.com/user-attachments/files/29498459/Muon.sin.txt)

Start by reproducing the slowdown with the attached Muon.sin.txt input at the reported 10M-event scale. Locate the Whizard execution and output path, then determine where processing effectively stops near 9M events. Done means the large sample completes without the reported slowdown and the result is verified against the reproduction case.

Written by the indexing model from the issue text.