Scalar evolution (SE or SCEV) models how values change inside loops. It can answer questions like “how many iterations does this loop perform?” and “What value does this variable have on the Nth iteration?” As such, it is the main building block for loop optimizations.
