[llvm] [DA] Rewrite BanerjeeMIV test with safe APInt interval arithmetic (PR #207662)

[Ryotaro Kasuga via llvm-commits]

kasuga-fj wrote:

The problem here is that the value of `WideBits` strongly depends on the subsequent computations. That is, whenever we change the logic, we may also need to reconsider its value; otherwise, we may miss the possibility of overflow someday. I don't think such a program is robust enough.

I also think that something like a dynamically-sized integer is not a particularly complex feature. Just replacing SCEV with it in the Banerjee MIV test might be a good first step.

https://github.com/llvm/llvm-project/pull/207662