feat: optimise BinningData, remove dead code

[asalzburger]

This PR cleans the BinUtility and BinningData objects:

• removes dead code
• removes hand-written search in vector in favour of std::lower_bound

It adds/removes relevant UnitTests and adds a new benchmark test, the benchmark before the changes shows:

16:36:17    BinUtility     INFO      Execution stats small: 1000 runs of 1 iteration(s), 0.0ms total, 0.0430+/-0.0001µs per run, 43.000+/-0.058ns per iteration
16:36:17    BinUtility     INFO      Fraction is: 21968011 vs. 21968011
16:36:19    BinUtility     INFO      Execution stats medium: 1000 runs of 1 iteration(s), 0.1ms total, 0.0510+/-0.0003µs per run, 51.000+/-0.346ns per iteration
16:36:19    BinUtility     INFO      Fraction is: 19641811 vs. 19641811
16:36:21    BinUtility     INFO      Execution stats many: 1000 runs of 1 iteration(s), 0.1ms total, 0.0780+/-0.0002µs per run, 78.000+/-0.173ns per iteration
16:36:21    BinUtility     INFO      Fraction is: 12780128 vs. 12780128
16:36:23    BinUtility     INFO      Execution stats many (lower_bounds): 1000 runs of 1 iteration(s), 0.0ms total, 0.0420+/-0.0001µs per run, 42.000+/-0.086ns per iteration
16:36:23    BinUtility     INFO      Fraction is: 23670634 vs. 23670634

to the new implementation:

17:25:25    BinUtility     INFO      Execution stats small: 1000 runs of 1 iteration(s), 0.0ms total, 0.0450+/-0.0001µs per run, 45.000+/-0.115ns per iteration
17:25:25    BinUtility     INFO      Fraction is: 21444179 vs. 21444179
17:25:27    BinUtility     INFO      Execution stats medium: 1000 runs of 1 iteration(s), 0.0ms total, 0.0460+/-0.0001µs per run, 46.000+/-0.115ns per iteration
17:25:27    BinUtility     INFO      Fraction is: 21030434 vs. 21030434
17:25:29    BinUtility     INFO      Execution stats many: 1000 runs of 1 iteration(s), 0.0ms total, 0.0480+/-0.0001µs per run, 48.000+/-0.115ns per iteration
17:25:29    BinUtility     INFO      Fraction is: 20028335 vs. 20028335
17:25:31    BinUtility     INFO      Execution stats equidistant: 1000 runs of 1 iteration(s), 0.0ms total, 0.0450+/-0.0001µs per run, 45.000+/-0.058ns per iteration
17:25:31    BinUtility     INFO      Fraction is: 20992746 vs. 20992746

The new implementation is faster in all bin scenarios.

Eventually, the BinUtility should go in favour of Grid usage, the removal of dead code is a first step towards this.

[codecov]

Codecov Report

Merging #657 (e494981) into master (bc366d9) will increase coverage by 0.02%.
The diff coverage is 50.00%.

@@            Coverage Diff             @@
##           master     #657      +/-   ##
==========================================
+ Coverage   49.04%   49.06%   +0.02%     
==========================================
  Files         329      329              
  Lines       16508    16438      -70     
  Branches     7692     7657      -35     
==========================================
- Hits         8097     8066      -31     
+ Misses       3028     2998      -30     
+ Partials     5383     5374       -9     

Impacted Files	Coverage Δ
Core/include/Acts/Utilities/BinUtility.hpp	45.83% <ø> (-0.96%)	⬇️
Core/include/Acts/Utilities/BinnedArray.hpp	33.33% <ø> (ø)
Core/include/Acts/Utilities/BinnedArrayXD.hpp	34.21% <ø> (+11.40%)	⬆️
Core/include/Acts/Utilities/BinningData.hpp	66.27% <50.00%> (+0.08%)	⬆️
Core/include/Acts/Propagator/Navigator.hpp	58.62% <0.00%> (ø)
Core/include/Acts/Propagator/EigenStepper.ipp	50.48% <0.00%> (+0.97%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update f3f5479...e494981. Read the comment docs.

[paulgessinger]

Seems fine to me.

[paulgessinger]

This changes the public API, but we probably wouldn't consider this breaking, right @asalzburger ?