Streamline constants

[valarauca]

Changes:

• Geodesic fields were are constants (not dependent on spheroid parameters) are now just constants.
• Fixed the type of many constants (i64 -> usize) where applicable.
• Remove arguments that are constants.
• Remove casts that were made redunant by changing i64 -> usize.
• Rewrite polyval; The new version is as close to optimal as my local

This is the first patch in a series of PR aiming at improving performance of geodesic inverse calculation.

• I agree to follow the project's code of conduct.
• I added an entry to CHANGES.md if knowledge of this change could be valuable to users.

---

[valarauca]

Minor changes to a few functions in geomath, sorry. Got a little distracted while copying & pasting stuff.

[valarauca]

A few nitpicks I spotted last night while looking this over.

[valarauca]

Aims to resolve:

• avoid passing GEODESIC_ORDER as argument #10
• favor type usize for variables that are primarily an array index #12
• replace get_epsilon with f64 EPSILON #1
• replace get_min_val with f64 MIN_POSITIVE #2
• replace geomath::cbrt with f64::cbrt #4
• replace geomath::fmod with direct calls to % operator #5

[michaelkirk]

There are some nice perf wins here, thanks!

$ cargo bench --bench="*" --  --baseline=main-2024-01-24 
   Compiling geographiclib-rs v0.2.3 (/Users/mkirk/src/georust/geographiclib-rs)
    Finished bench [optimized] target(s) in 1.03s
     Running benches/geodesic_benchmark.rs (target/release/deps/geodesic_benchmark-656e1aa161ec181c)
direct (c wrapper)/default
                        time:   [23.868 µs 23.894 µs 23.923 µs]
                        change: [-0.1047% +0.2586% +0.8488%] (p = 0.32 > 0.05)
                        No change in performance detected.
Found 6 outliers among 100 measurements (6.00%)
  3 (3.00%) high mild
  3 (3.00%) high severe

direct (rust impl)/default
                        time:   [26.527 µs 26.565 µs 26.609 µs]
                        change: [-11.679% -11.528% -11.381%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 8 outliers among 100 measurements (8.00%)
  1 (1.00%) low mild
  3 (3.00%) high mild
  4 (4.00%) high severe

inverse (c wrapper)/default
                        time:   [44.686 µs 44.742 µs 44.804 µs]
                        change: [-2.2765% -0.7738% +0.2501%] (p = 0.30 > 0.05)
                        No change in performance detected.
Found 6 outliers among 100 measurements (6.00%)
  2 (2.00%) high mild
  4 (4.00%) high severe

inverse (rust impl)/default
                        time:   [58.362 µs 58.459 µs 58.567 µs]
                        change: [-21.016% -20.859% -20.699%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 3 outliers among 100 measurements (3.00%)
  2 (2.00%) high mild
  1 (1.00%) high severe

This PR entails a lot of different changes. Would you be willing to break it up a bit to be easier to review?

[valarauca]

Would you be willing to break it up a bit to be easier to review?

I already did or I hope I did? 😅 my eventual goal was to wind up something like this.

I guess I can try to subdivide this further. I am doing 2 different things in 1 PR (standardizing constants & changing i64 to usize).

Let me see if I can reduce the scope slightly