MadNLP / DynamicNLPModels.jl

NLPModels for dynamic optimization

MIT License

11 stars 1 forks source link

Increased efficiency of `add_jtsj!` and `mul!` #36

Closed dlcole3 closed 2 years ago

dlcole3 commented 2 years ago

Changed the order of calculations within the for loops of add_jtsj! so that $\Sigma B$ was calculated fewer times. Doing so improved the overall performance of the function so that add_jtsj! is now at least marginally faster than mul! when scaling with N, ns, or nu.

Also updated mul! extensions so that they use one less loop. They now operate on longer blocks than before.

codecov-commenter commented 2 years ago

Codecov Report

Merging #36 (cbfabd5) into main (1ae317b) will decrease coverage by 0.00%. The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main      #36      +/-   ##
==========================================
- Coverage   97.56%   97.56%   -0.01%     
==========================================
  Files           1        1              
  Lines        1068     1066       -2     
==========================================
- Hits         1042     1040       -2     
  Misses         26       26

Impacted Files	Coverage Δ
src/DynamicNLPModels.jl	`97.56% <100.00%> (-0.01%)`	:arrow_down: