Convert Box to BoxND

[AlexanderSinn]

Summary

Similar to #3969 and #3988 but for Box.

Additional background

It should be checked that the changes to BoxIndexer do not affect the compiled GPU code. In my testing, it gives the same performance as development. 

Example usage:

amrex::BoxND b1{amrex::IntVectND{1,2,3}, amrex::IntVectND{4,5,6}, amrex::IntVectND{1,0,1}};
// ((1,2,3) (4,5,6) (1,0,1))
auto b2 = amrex::BoxCat(b1, b1, b1);
// ((1,2,3,1,2,3,1,2,3) (4,5,6,4,5,6,4,5,6) (1,0,1,1,0,1,1,0,1))
auto [b3, b4, b5, b6, b7] = amrex::BoxSplit<1, 4, 2, 1, 1>(b2);
// ((1) (4) (1))((2,3,1,2) (5,6,4,5) (0,1,1,0))((3,1) (6,4) (1,1))((2) (5) (0))((3) (6) (1))
auto b8 = amrex::BoxResize<2>(b4);
// ((2,3) (5,6) (0,1))
auto b9 = amrex::BoxResize<5>(b8);
// ((2,3,0,0,0) (5,6,0,0,0) (0,1,0,0,0))

Checklist

The proposed changes:

• [ ] fix a bug or incorrect behavior in AMReX
• [x] add new capabilities to AMReX
• [ ] changes answers in the test suite to more than roundoff level
• [ ] are likely to significantly affect the results of downstream AMReX users
• [ ] include documentation in the code and/or rst files, if appropriate

[AlexanderSinn]

I added IntVectND::dim3(int fill_extra) to make it easier to implement BoxND functions such as end and length. It looks like pyamrex needs an overload_cast for dim3() now.

PR: https://github.com/AMReX-Codes/pyamrex/pull/345

[ax3l]

https://github.com/AMReX-Codes/pyamrex/pull/345 merged - thanks!

[WeiqunZhang]

LGTM. I will do one more pass tomorrow.