We need to use long long for particle IDs to support more than 2 billion (2^31) particles. But there is no reason to support having so many particles in a single MPI process. There is a potential cost to using long long in a particle loop. Therefore, I am changing to use int, preceeded by an assertion that nop is no greater than INT_MAX, in the following two critical methods:
We need to use
long long
for particle IDs to support more than 2 billion (2^31) particles. But there is no reason to support having so many particles in a single MPI process. There is a potential cost to using long long in a particle loop. Therefore, I am changing to use int, preceeded by an assertion that nop is no greater than INT_MAX, in the following two critical methods: