refactor!: Reduce heap allocations in navigation

[paulgessinger]

This PR changes the storage values in Navigator to use boost::container::small_vector for layers, sensitive and boundary surfaces. At the same time Layer::compatibleSurface, TrackingVolume::combatibleBoundaries and TrackingVolume::compatibleLayers are modified to return this container.

This PR contains the changes from #1119, so should be merged afterwards.

I benchmarked this with the EigenStepper without covariance transport, as this shouldn't be affected by this change.

On main I get:

  Time (mean ± σ):     14.518 s ±  0.222 s    [User: 882.433 s, System: 0.663 s]
  Range (min … max):   14.249 s … 14.836 s    5 runs

while on this branch I get

  Time (mean ± σ):     13.981 s ±  0.141 s    [User: 849.680 s, System: 0.482 s]
  Range (min … max):   13.879 s … 14.228 s    5 runs

for 1000 tracks per event and 20000 events each.

The measurements seem to be just outside about 2sigma, if we take the mean at face-value, it goes down by about 3%.

[paulgessinger]

@asalzburger what do you think?

[codecov]

Codecov Report

Merging #1124 (25b70ed) into main (b6695fd) will decrease coverage by 0.03%.
The diff coverage is 60.00%.

@@            Coverage Diff             @@
##             main    #1124      +/-   ##
==========================================
- Coverage   48.16%   48.13%   -0.04%     
==========================================
  Files         341      341              
  Lines       17646    17660      +14     
  Branches     8327     8342      +15     
==========================================
+ Hits         8499     8500       +1     
  Misses       3372     3372              
- Partials     5775     5788      +13     

Impacted Files	Coverage Δ
Core/include/Acts/Geometry/Layer.hpp	100.00% <ø> (ø)
Core/include/Acts/Geometry/TrackingVolume.hpp	69.23% <ø> (ø)
Core/include/Acts/Propagator/Navigator.hpp	55.11% <50.00%> (-1.65%)	⬇️
Core/src/Surfaces/TrapezoidBounds.cpp	75.75% <56.25%> (-13.14%)	⬇️
Core/src/Geometry/Layer.cpp	57.57% <100.00%> (+0.75%)	⬆️
Core/src/Geometry/TrackingVolume.cpp	42.99% <100.00%> (-0.37%)	⬇️

---

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 a0e4a18...25b70ed. Read the comment docs.

[HadrienG2]

@paulgessinger Try re-running the benchmark with more runs. "just outside 2 sigmas" on 5 runs is not enough to claim a significant difference on benchmark timings, whose probability law is not normal.

[paulgessinger]

Sure. This is what I get with 30 runs on main

  Time (mean ± σ):     14.444 s ±  0.183 s    [User: 878.311 s, System: 0.630 s]
  Range (min … max):   14.141 s … 14.867 s    30 runs

and on this PR

  Time (mean ± σ):     13.977 s ±  0.235 s    [User: 848.936 s, System: 0.520 s]
  Range (min … max):   13.653 s … 14.603 s    30 runs

The means seem roughly consistent with before. I can run again with even more samples.

Overall, this isn't a huge performance improvement, I agree. But looking at the propagation profile, at least I can't see any great singular optimization opportunities, where we could get a large increase in performance in one go. It rather seems like there's a number of smaller optimization opportunities that in total could be a net positive.

[robertlangenberg]

The changes look good, can you resolve the conflicts in Layer.hpp and Layer.cpp?