Tags: SciML/NonlinearSolve.jl
Tags
[Diff since BracketingNonlinearSolve-v1.2.0](BracketingNonlinearSolve… …-v1.2.0...BracketingNonlinearSolve-v1.3.0) **Merged pull requests:** - Add a GPU tutorial (#589) (@ChrisRackauckas) - Add Alefeld method to documentation page (#591) (@fgittins) - chore(deps): bump julia-actions/julia-format from 3 to 4 (#592) (@dependabot[bot]) - Update CI_NonlinearSolve.yml (#594) (@ChrisRackauckas) - Fix test error for non-CUDA machines (#595) (@fgittins) - early exit in SimpleTrustRegion (#598) (@ChrisRackauckas) - CompatHelper: bump compat for SciMLOperators to 0.4 for package NonlinearSolveBase, (keep existing compat) (#601) (@github-actions[bot]) - CompatHelper: bump compat for SciMLOperators to 0.4 for package NonlinearSolveQuasiNewton, (keep existing compat) (#602) (@github-actions[bot]) - CompatHelper: bump compat for SciMLOperators to 0.4 for package SciMLJacobianOperators, (keep existing compat) (#603) (@github-actions[bot]) - Handle default solver for SCCs (#604) (@ChrisRackauckas) - Fix generic fallbacks for L2Norm (#605) (@ChrisRackauckas) - Fix return codes for nonlinear least squares (#606) (@ChrisRackauckas) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package SciMLJacobianOperators, (keep existing compat) (#607) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package SimpleNonlinearSolve, (keep existing compat) (#608) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7, (keep existing compat) (#609) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package docs, (keep existing compat) (#610) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package NonlinearSolveBase, (keep existing compat) (#611) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package NonlinearSolveHomotopyContinuation, (keep existing compat) (#612) (@github-actions[bot]) - Bump to use SciMLBase ImmutableNonlinearProblem (#619) (@ChrisRackauckas) - Continue on with the polyalg if a stalledsuccess (#622) (@ChrisRackauckas) - Implement Adjoints for solution of IntervalNonlinearProblems (#623) (@jClugstor) - Fix type instability in LinearSolve QR factorization fallback (#624) (@ChrisRackauckas) - Bump SciMLOperators v1 (#625) (@ChrisRackauckas) - Set strict=false for DI in HomotopyContinuation (#626) (@ChrisRackauckas) - Make Aqua happy (#628) (@ChrisRackauckas) **Closed issues:** - Add a tutorial on using SimpleNonlinearSolve inside kernels (#405) - NonlinearLeastSquaresProblem returns stalled (#459) - Set of benchmarks to optimize AD? (#479) - Release Checklist for `v4` (#481) - L2_NORM resorts to scalar indexing on GPU ComponentArray (#524) - `IntervalNonlinearProblem` fails with `Zygote` (#581) - `Muller` algorithm benchmark (#582) - Solver fail when jac_prototype is BandedMatrix{BigFloat} (#587) - `SimpleTrustRegion` fails for SCC decomposition of problem (#596) - Levenberg-Marquardt with Sparse J: J isn't square?! (#599) - Can't update packages that have JuliaSimCompiler in their dependencies (#617)
[Diff since SciMLJacobianOperators-v0.1.5](SciMLJacobianOperators-v0.… …1.5...SciMLJacobianOperators-v0.1.6) **Merged pull requests:** - Fix return codes for nonlinear least squares (#606) (@ChrisRackauckas) - Bump to use SciMLBase ImmutableNonlinearProblem (#619) (@ChrisRackauckas) - Continue on with the polyalg if a stalledsuccess (#622) (@ChrisRackauckas) - Implement Adjoints for solution of IntervalNonlinearProblems (#623) (@jClugstor) - Bump SciMLOperators v1 (#625) (@ChrisRackauckas) **Closed issues:** - NonlinearLeastSquaresProblem returns stalled (#459) - Set of benchmarks to optimize AD? (#479) - Release Checklist for `v4` (#481) - `Muller` algorithm benchmark (#582) - Can't update packages that have JuliaSimCompiler in their dependencies (#617)
[Diff since NonlinearSolveQuasiNewton-v1.5.0](NonlinearSolveQuasiNewt… …on-v1.5.0...NonlinearSolveQuasiNewton-v1.6.0) **Merged pull requests:** - Fix return codes for nonlinear least squares (#606) (@ChrisRackauckas) - Bump to use SciMLBase ImmutableNonlinearProblem (#619) (@ChrisRackauckas) - Continue on with the polyalg if a stalledsuccess (#622) (@ChrisRackauckas) - Implement Adjoints for solution of IntervalNonlinearProblems (#623) (@jClugstor) - Bump SciMLOperators v1 (#625) (@ChrisRackauckas) **Closed issues:** - NonlinearLeastSquaresProblem returns stalled (#459) - Set of benchmarks to optimize AD? (#479) - Release Checklist for `v4` (#481) - `Muller` algorithm benchmark (#582) - Can't update packages that have JuliaSimCompiler in their dependencies (#617)
[Diff since NonlinearSolveBase-v1.11.0](NonlinearSolveBase-v1.11.0...… …NonlinearSolveBase-v1.12.0) **Merged pull requests:** - Fix type instability in LinearSolve QR factorization fallback (#624) (@ChrisRackauckas) - Set strict=false for DI in HomotopyContinuation (#626) (@ChrisRackauckas)
[Diff since NonlinearSolveBase-v1.10.0](NonlinearSolveBase-v1.10.0...… …NonlinearSolveBase-v1.11.0) **Merged pull requests:** - Implement Adjoints for solution of IntervalNonlinearProblems (#623) (@jClugstor) - Bump SciMLOperators v1 (#625) (@ChrisRackauckas)
[Diff since v4.8.0](v4.8.0...v4.9.0) **Merged pull requests:** - Add a GPU tutorial (#589) (@ChrisRackauckas) - Add Alefeld method to documentation page (#591) (@fgittins) - chore(deps): bump julia-actions/julia-format from 3 to 4 (#592) (@dependabot[bot]) - Update CI_NonlinearSolve.yml (#594) (@ChrisRackauckas) - Fix test error for non-CUDA machines (#595) (@fgittins) - early exit in SimpleTrustRegion (#598) (@ChrisRackauckas) - CompatHelper: bump compat for SciMLOperators to 0.4 for package NonlinearSolveBase, (keep existing compat) (#601) (@github-actions[bot]) - CompatHelper: bump compat for SciMLOperators to 0.4 for package NonlinearSolveQuasiNewton, (keep existing compat) (#602) (@github-actions[bot]) - CompatHelper: bump compat for SciMLOperators to 0.4 for package SciMLJacobianOperators, (keep existing compat) (#603) (@github-actions[bot]) - Handle default solver for SCCs (#604) (@ChrisRackauckas) - Fix generic fallbacks for L2Norm (#605) (@ChrisRackauckas) - Fix return codes for nonlinear least squares (#606) (@ChrisRackauckas) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package SciMLJacobianOperators, (keep existing compat) (#607) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package SimpleNonlinearSolve, (keep existing compat) (#608) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7, (keep existing compat) (#609) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package docs, (keep existing compat) (#610) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package NonlinearSolveBase, (keep existing compat) (#611) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package NonlinearSolveHomotopyContinuation, (keep existing compat) (#612) (@github-actions[bot]) - Bump to use SciMLBase ImmutableNonlinearProblem (#619) (@ChrisRackauckas) - Continue on with the polyalg if a stalledsuccess (#622) (@ChrisRackauckas) **Closed issues:** - Add a tutorial on using SimpleNonlinearSolve inside kernels (#405) - NonlinearLeastSquaresProblem returns stalled (#459) - Set of benchmarks to optimize AD? (#479) - Release Checklist for `v4` (#481) - L2_NORM resorts to scalar indexing on GPU ComponentArray (#524) - `Muller` algorithm benchmark (#582) - Solver fail when jac_prototype is BandedMatrix{BigFloat} (#587) - `SimpleTrustRegion` fails for SCC decomposition of problem (#596) - Levenberg-Marquardt with Sparse J: J isn't square?! (#599) - Can't update packages that have JuliaSimCompiler in their dependencies (#617)
[Diff since NonlinearSolveBase-v1.9.0](NonlinearSolveBase-v1.9.0...No… …nlinearSolveBase-v1.10.0) **Merged pull requests:** - Bump to use SciMLBase ImmutableNonlinearProblem (#619) (@ChrisRackauckas) - Continue on with the polyalg if a stalledsuccess (#622) (@ChrisRackauckas) **Closed issues:** - Set of benchmarks to optimize AD? (#479) - Release Checklist for `v4` (#481) - `Muller` algorithm benchmark (#582) - Can't update packages that have JuliaSimCompiler in their dependencies (#617)
[Diff since v3.15.1](v3.15.1...v3.15.2) This release has been identified as a backport. Automated changelogs for backports tend to be wildly incorrect. Therefore, the list of issues and pull requests is hidden. <!-- **Merged pull requests:** - feat: `NonlinearSolveBase` + `BracketingNonlinearSolve` + `SimpleNonlinearSolve` (#458) (@avik-pal) - fix: remove uses of LineSearches.jl backtracking (#480) (@avik-pal) - feat: add `PETScSNES` (#482) (@avik-pal) - refactor: splitting up of NonlinearSolve (#483) (@avik-pal) - refactor!: move preconditioners inside linear solvers (#485) (@avik-pal) - fix: remove the common files (#486) (@avik-pal) - fix: restore the history of SimpleNonlinearSolve (#487) (@avik-pal) - fix: hessian (#489) (@avik-pal) - ci: run windows tests with a single worker (#490) (@avik-pal) - fix: reinit! on forwarddiff cache (#491) (@avik-pal) - refactor: Move NonlinearSolvePolyAlgorithm to Base (#494) (@ErikQQY) - fix: `PolyesterForwardDiff` shouldn't be the default for static arrays (#495) (@avik-pal) - Fix ambiguity in `__solve` with polyalgorithm due to dispatch on `Va` (#498) (@ChrisRackauckas) - refactor: Move RobustMultiNewton to NonlinearSolveFirstOrder (#499) (@ErikQQY) - fix: fix SII implementation for caches, tests (#500) (@AayushSabharwal) - Create SCCNonlinearSolve (#502) (@ChrisRackauckas) - chore(deps): bump codecov/codecov-action from 4 to 5 (#505) (@dependabot[bot]) - fix: fix SII implementation for `NonlinearSolveNoInitCache` (#506) (@AayushSabharwal) - Update release_notes.md (#508) (@briochemc) - fix: fix MTK indexing tests (#509) (@AayushSabharwal) - refactor: Move dual nonlinear solving to NonlinearSolveBase (#513) (@ErikQQY) - fix: remove compiled reversediff from default options (#514) (@avik-pal) - fix: enzyme is now properly supported in 1.11 (#515) (@avik-pal) - feat: add hooks for `OverrideInit` (#517) (@AayushSabharwal) - build: bump versions of `NonlinearSolve` and sub-packages (#518) (@AayushSabharwal) - Fix incorrect Base version (#520) (@ErikQQY) - feat: add `NonlinearSolveHomotopyContinuation.jl` (#523) (@AayushSabharwal) - docs: Fix typo (#527) (@mrkn) - feat(HomotopyContinuation): enable more performant jacobians with Enzyme (#528) (@AayushSabharwal) - fix(HomotopyContinuation): handle `NaN` results from `unpolynomialize` (#529) (@AayushSabharwal) - Bump `LinearSolve` compat version in `NonlinearSolveBase` (#531) (@ytdHuang) - Use aliasing API for linear solves (#533) (@jClugstor) - fix type stability of SimpleNonlinearSolve (#536) (@oscardssmith) - Fix markdown list in faq.md (#538) (@abhro) - Update compats (#539) (@ChrisRackauckas) - ci(CompatHelper): checkout the repository (#542) (@thazhemadam) - CompatHelper: bump compat for FixedPointAcceleration in [weakdeps] to 1, (keep existing compat) (#543) (@github-actions[bot]) - CompatHelper: add new compat entry for NonlinearSolveHomotopyContinuation at version 0.1 for package docs, (keep existing compat) (#544) (@github-actions[bot]) - CompatHelper: bump compat for LinearSolve to 3 for package docs, (keep existing compat) (#545) (@github-actions[bot]) - Add misc test group to NonlinearSolve CI to run QA tests (#546) (@ChrisRackauckas) - minor itp speedups (#547) (@oscardssmith) - more ITP simplification (#548) (@oscardssmith) - Migrate the CUDA Termination Conditions tests from DiffEqBase (#549) (@ChrisRackauckas) - More generic safe similar (#550) (@ErikQQY) - CompatHelper: bump compat for AlgebraicMultigrid to 1 for package docs, (keep existing compat) (#551) (@github-actions[bot]) - support ForwardDiff 1.0 (#561) (@oscardssmith) - Update Project.toml (#562) (@ChrisRackauckas) - Fix ForwardDiff overload for user vjp (#563) (@ChrisRackauckas) - support Zygote@0.7 (#566) (@oscardssmith) - Add Muller's method (#568) (@fgittins) - Fix SCCNonlinearSolve test error (#569) (@fgittins) - feat: use sources (#571) (@ErikQQY) - Fix CI (#572) (@ChrisRackauckas) - Add documentation for Muller's method (#573) (@fgittins) - remove true_jacobian Broyden from default alg (#577) (@oscardssmith) - fix: initial objective for termination conditions (#579) (@avik-pal) - fix: failing tests (#580) (@avik-pal) - Simplify requirements from default polyalg (#583) (@ChrisRackauckas) - Fix NonlinearSolveFirstOrder type definition order (#584) (@ChrisRackauckas) - Update 23_test_problems_tests.jl (#585) (@ChrisRackauckas) - Add a GPU tutorial (#589) (@ChrisRackauckas) - Add Alefeld method to documentation page (#591) (@fgittins) - chore(deps): bump julia-actions/julia-format from 3 to 4 (#592) (@dependabot[bot]) - Update CI_NonlinearSolve.yml (#594) (@ChrisRackauckas) - Fix test error for non-CUDA machines (#595) (@fgittins) - early exit in SimpleTrustRegion (#598) (@ChrisRackauckas) - CompatHelper: bump compat for SciMLOperators to 0.4 for package NonlinearSolveBase, (keep existing compat) (#601) (@github-actions[bot]) - CompatHelper: bump compat for SciMLOperators to 0.4 for package NonlinearSolveQuasiNewton, (keep existing compat) (#602) (@github-actions[bot]) - CompatHelper: bump compat for SciMLOperators to 0.4 for package SciMLJacobianOperators, (keep existing compat) (#603) (@github-actions[bot]) - Handle default solver for SCCs (#604) (@ChrisRackauckas) - Fix generic fallbacks for L2Norm (#605) (@ChrisRackauckas) - Fix return codes for nonlinear least squares (#606) (@ChrisRackauckas) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package SciMLJacobianOperators, (keep existing compat) (#607) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package SimpleNonlinearSolve, (keep existing compat) (#608) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7, (keep existing compat) (#609) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package docs, (keep existing compat) (#610) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package NonlinearSolveBase, (keep existing compat) (#611) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package NonlinearSolveHomotopyContinuation, (keep existing compat) (#612) (@github-actions[bot]) **Closed issues:** - An interface to PETSc SNES nonlinear solve would be nice (#160) - Use LineSearch.jl directly (once it is ready) (#281) - Default Automatic Differentiation Choice (#353) - Strongly Connected Component (SCC) Scheduled Nonlinear Problems (#388) - `reinit` cache for Forward Mode AD not working (#391) - Muller's method (#403) - Add a tutorial on using SimpleNonlinearSolve inside kernels (#405) - MethodError: anyeltypedual(::Type{Union{}}) when passing ODEProblem as a parameter (#440) - Implementing ForwardDiff.hessian through NonlinearLeastSquaresProblem (#445) - [RFC] Splitting up NonlinearSolve (#456) - NonlinearLeastSquaresProblem returns stalled (#459) - New Jacobian cache causes copy bug (#473) - Add direct Enzyme support (#476) - Unexpected `MaxIters` retcode (#492) - Failed to solve the set of equations - Potential Rank Deficient Matrix and ReturnCode.Stalled = 17 (#496) - Ambiguity error when solving NonlinearProblem on master (#497) - ITP implementation beaten by Roots.jl (#501) - Add TagBot for Subpackages (#503) - Circular dependency precompilation hangs (#511) - Access to undefined reference when solving problems with Dual of BigFloat (#512) - Julia 1.11 crashes with NonlinearSolve 4 when ReverseDiff is loaded (#516) - latest release fails to precompile w/ OrdinaryDiffEq.jl (#519) - Unable to `init on `IntervalNonlinearProblem` (#521) - L2_NORM resorts to scalar indexing on GPU ComponentArray (#524) - Tutorial example fails (#526) - Use `NonlinearAliasSpecifier` to specify aliasing during solve (#552) - Muller's method (#564) - NonlinearLeastSquares does not converge to or accept the correct minimum (#565) - CI warnings on 1.12 prerelease (#576) - Solver fail when jac_prototype is BandedMatrix{BigFloat} (#587) - `SimpleTrustRegion` fails for SCC decomposition of problem (#596) - Levenberg-Marquardt with Sparse J: J isn't square?! (#599) -->
[Diff since NonlinearSolveBase-v1.8.0](NonlinearSolveBase-v1.8.0...No… …nlinearSolveBase-v1.9.0) **Merged pull requests:** - Fix return codes for nonlinear least squares (#606) (@ChrisRackauckas) **Closed issues:** - NonlinearLeastSquaresProblem returns stalled (#459)
[Diff since SimpleNonlinearSolve-v2.4.0](SimpleNonlinearSolve-v2.4.0.… …..SimpleNonlinearSolve-v2.5.0) **Merged pull requests:** - CompatHelper: bump compat for SciMLOperators to 0.4 for package NonlinearSolveBase, (keep existing compat) (#601) (@github-actions[bot]) - CompatHelper: bump compat for SciMLOperators to 0.4 for package NonlinearSolveQuasiNewton, (keep existing compat) (#602) (@github-actions[bot]) - CompatHelper: bump compat for SciMLOperators to 0.4 for package SciMLJacobianOperators, (keep existing compat) (#603) (@github-actions[bot]) - Handle default solver for SCCs (#604) (@ChrisRackauckas) - Fix generic fallbacks for L2Norm (#605) (@ChrisRackauckas) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package SciMLJacobianOperators, (keep existing compat) (#607) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package SimpleNonlinearSolve, (keep existing compat) (#608) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7, (keep existing compat) (#609) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package docs, (keep existing compat) (#610) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package NonlinearSolveBase, (keep existing compat) (#611) (@github-actions[bot]) - CompatHelper: bump compat for DifferentiationInterface to 0.7 for package NonlinearSolveHomotopyContinuation, (keep existing compat) (#612) (@github-actions[bot]) **Closed issues:** - L2_NORM resorts to scalar indexing on GPU ComponentArray (#524) - Levenberg-Marquardt with Sparse J: J isn't square?! (#599)
PreviousNext