Skip to content

NonlinearSolveAlg: let inner solver refresh J when no W is reused (fixes master nlstep order failures) #3818

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
ChrisRackauckas-Claude wants to merge 1 commit into
base: master
Choose a base branch
from
+14 −2
Draft

NonlinearSolveAlg: let inner solver refresh J when no W is reused (fixes master nlstep order failures) #3818

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 14 additions & 2 deletions lib/OrdinaryDiffEqNonlinearSolve/src/newton.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -145,7 +145,13 @@ end
(; tstep, invγdt) = cache

nlcache = nlsolver.cache.cache
recompute_jacobian = nlsolver.iter == 1 && (cache.W === nothing || cache.new_W)
# When no reusable W exists (nlstep/DAE/OOP), the inner NonlinearSolve must be
# allowed to refresh its Jacobian every outer iteration: freezing it after iter 1
# degrades the inner solve to modified Newton while the outer η·ndz < κ test still
# assumes its calibration, leaving a dt-independent unconverged remainder per step
# (FBDF nlstep convergence order collapsed to ~0.6). With a reused W the Jacobian
# is the ODE-side W by construction, so refreshing beyond iter 1 is pointless.
recompute_jacobian = cache.W === nothing || (nlsolver.iter == 1 && cache.new_W)
step!(nlcache; recompute_jacobian)
nlsolver.ztmp = nlcache.u

Expand All @@ -170,7 +176,13 @@ end

nlstep_data = integrator.f.nlstep_data
nlcache = nlsolver.cache.cache
recompute_jacobian = nlsolver.iter == 1 && (cache.W === nothing || cache.new_W)
# When no reusable W exists (nlstep/DAE/OOP), the inner NonlinearSolve must be
# allowed to refresh its Jacobian every outer iteration: freezing it after iter 1
# degrades the inner solve to modified Newton while the outer η·ndz < κ test still
# assumes its calibration, leaving a dt-independent unconverged remainder per step
# (FBDF nlstep convergence order collapsed to ~0.6). With a reused W the Jacobian
# is the ODE-side W by construction, so refreshing beyond iter 1 is pointless.
recompute_jacobian = cache.W === nothing || (nlsolver.iter == 1 && cache.new_W)
step!(nlcache; recompute_jacobian)

if nlstep_data !== nothing
Expand Down
Loading