Speeding up subsolvers

[mpetrosian]

• render bntr subsolver numba-compatible. This includes jitting the following functions:
  • take_preliminary_gradient_descent_step_and_check_for_solution
  • compute_conjugate_gradient_step
  • minimize_trust_trsbox
  • minimize_trust_cg
  • minimize_trust_stcg

[codecov]

Codecov Report

Merging #391 (c9c23d2) into main (a7c0232) will increase coverage by 0.08%.
The diff coverage is 99.78%.

@@            Coverage Diff             @@
##             main     #391      +/-   ##
==========================================
+ Coverage   92.78%   92.87%   +0.08%     
==========================================
  Files         209      211       +2     
  Lines       16628    17070     +442     
==========================================
+ Hits        15428    15853     +425     
- Misses       1200     1217      +17     

Impacted Files	Coverage Δ
...ptimization/subsolvers/bounded_newton_quadratic.py	86.40% <ø> (ø)
...imagic/optimization/subsolvers/gqtpar_quadratic.py	89.31% <0.00%> (ø)
...magic/optimization/subsolvers/_trsbox_quadratic.py	85.36% <100.00%> (-6.88%)	⬇️
...ic/optimization/subsolvers/quadratic_subsolvers.py	96.29% <100.00%> (+0.46%)	⬆️
...timagic/optimization/tranquilo/solve_subproblem.py	88.46% <100.00%> (+0.14%)	⬆️
tests/optimization/tranquilo/test_bntr_fast.py	100.00% <100.00%> (ø)
...timization/tranquilo/test_minimize_trust_region.py	100.00% <100.00%> (ø)
...ts/optimization/tranquilo/test_solve_subproblem.py	100.00% <100.00%> (ø)

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[janosg]

Ran another set of benchmarks. If anything, the faster subsolvers also perform better.