In HierarchicalLSCM::solveLSCMLevel, the face assembly loop calls std::sin and std::cos per edge on each face. These are computed inside the loop body rather than precomputed into a flat array before the loop. Precomputing would allow the compiler to vectorize the loop (sin/cos calls prevent auto-vectorization) and reduces redundant transcendental evaluations when an edge's angle appears in multiple adjacent faces.
Identified during code review of PR #44 (low priority).
In
HierarchicalLSCM::solveLSCMLevel, the face assembly loop callsstd::sinandstd::cosper edge on each face. These are computed inside the loop body rather than precomputed into a flat array before the loop. Precomputing would allow the compiler to vectorize the loop (sin/cos calls prevent auto-vectorization) and reduces redundant transcendental evaluations when an edge's angle appears in multiple adjacent faces.Identified during code review of PR #44 (low priority).