ODD ttbar Clusterization Fix, main branch (2024.05.04.)

[krasznaa]

Made high-$\mu$ ODD ttbar samples succeed with (device) clusterization.

As discussed in #565, with $\mu$ = 100 and higher, on the files that I generated as described in #561, I had to do this to avoid running into assertion errors from the clusterization code. 🤔

I'm very open to discussions about why this is, but the code, with its current logic of how it splits the cells into partitions, does need this update for the highest pileup ODD simulations.

[stephenswat]

This is, of course, a perfectly fine change to make. Just note that the trade-off here is that increasing this parameter is expected to decrease occupancy and, as a result, performance. But that's something we can figure out later.