Yes, division is more “expensive” in a couple of ways, neither of which matter in this case.
First, division has more latency than multiplication (so the CPU has to wait longer for the result). The compiler can get around that by scheduling other work in the meantime unless the result is needed immediately for the next computation.
Second, the CPU (depending on type) might be able to do more multiplication operations in parallel, because more silicon is allocated to this operation (which is used more often than division). However, this only matters if enough work can be supplied in parallel.
The operation that has the lowest throughput in this particular case is probably the memory read/writes, so the difference between multiplication and division “cost” is not important.