> Do you think it’s possible to calculate the relationship between the depth of the git data and the number of I/O operations
It's actually pretty straight forward since you know exactly what operations has to be done. With very good I/O, you may find the CPU will become your bottleneck.