v6.77.0

OrdinaryDiffEq v6.77.0

Diff since v6.76.0

Merged pull requests:

• shorten timespan on initdt test (#2194) (@oscardssmith)
• I think this fixes a few minor bugs with nonlinear convergence (#2196) (@oscardssmith)
• Fix and test DPRKN interpolation with idxs (#2201) (@ChrisRackauckas)

v6.76.0

OrdinaryDiffEq v6.76.0

Diff since v6.75.0

Merged pull requests:

• Support NonlinearSolve.jl for the ODE nonlinear solver (#2167) (@oscardssmith)
• Adding Pseudo-Symplectic Explicit Runge-Kutta methods (#2174) (@MarcoArtiano)
• added step_limiter to RADAUIIA3,5 and made basic tests for step_limiter (#2176) (@cwittens)
• test(inplace_interpolation.jl): check for approx equality (#2181) (@characat0)
• refactor: use tunable parameters for cache in dae initialization (#2186) (@sathvikbhagavan)
• added step_limiter to more methods (#2190) (@cwittens)
• CompatHelper: add new compat entry for SciMLStructures at version 1, (keep existing compat) (#2191) (@github-actions[bot])
• fix dependent mass matrix tests (#2192) (@ChrisRackauckas)
• Fix relaxation test (#2193) (@ChrisRackauckas)

v6.75.0

OrdinaryDiffEq v6.75.0

Diff since v6.74.1

Merged pull requests:

• Implement 4th order Runge-Kutta-Nystrom method (#2161) (@Shreyas-Ekanathan)
• added ROS2 method and added ROS2, ROS23 and ROS34PW methods to iipvsoop_test.jl (#2165) (@cwittens)
• Added Runge–Kutta pairs of orders 9(8) for use in quadruple precision computations (#2166) (@Theozeud)
• Bump julia-actions/setup-julia from 1 to 2 (#2169) (@dependabot[bot])
• Modifications Rodas5Pe, Rodas5Pr of Rodas5P concerning issue #2054 and minor bug fixes of Rodas3P/23W (#2170) (@gstein3m)
• CompatHelper: bump compat for ADTypes to 1, (keep existing compat) (#2172) (@github-actions[bot])
• added algos to rosenbrock.md (#2175) (@cwittens)
• Fix issue #2173 divide-by-zero error (#2178) (@afossa)
• Bump version and release for downstream CI with ADTypes 1.0 (#2180) (@Vaibhavdixit02)

Closed issues:

• Investigate the possibility of StaticArrays tableau methods (#114)
• Analysis of time spent precompiling (#1937)
• Second-order Rosenbrock method of Verwer, Spee, Blom, Hundsdorfer (1999) (#2112)
• New Runge–Kutta pairs of orders 9(8) for use in quadruple precision computations (#2159)
• Plotting analytic solutions fails (#2162)
• Divide-by-zero error thrown during vector initialization (#2173)
• unstable_check gets passed value that's far from solution (#2179)

v6.74.1

OrdinaryDiffEq v6.74.1

Diff since v6.74.0

Merged pull requests:

• Avoid throwing error if type inference on u0/t fails (#2157) (@mchitre)
• fix DAE init when BrownFullBasic is given a specific nlsolver to use (#2160) (@oscardssmith)

Closed issues:

• Automatically chosen algorithm by solve(prob; alg_hints=:auto) fails to solve ODEs (#2146)

v6.74.0

OrdinaryDiffEq v6.74.0

Diff since v6.73.1

Merged pull requests:

• Bump codecov/codecov-action from 3 to 4 (#2129) (@dependabot[bot])
• Directly enforce algebraic condition correctness in ImplicitEuler (#2153) (@ChrisRackauckas)
• Reset uprev with reinitializations (#2155) (@ChrisRackauckas)

v6.73.1

OrdinaryDiffEq v6.73.1

Diff since v6.73.0

v6.73.0

OrdinaryDiffEq v6.73.0

Diff since v6.72.0

Merged pull requests:

• Allow for DAE initialization on ODEs with initializeprob (#2154) (@ChrisRackauckas)

v6.72.0

OrdinaryDiffEq v6.72.0

Diff since v6.71.0

Merged pull requests:

• Improve mass matrix error message (#2145) (@ChrisRackauckas)
• fix memory leak on repeated init(prob) (#2148) (@oscardssmith)
• reapply formatter (#2150) (@ArnoStrouwen)
• Allow for using the initializeprob with OverrideInit (#2151) (@ChrisRackauckas)

Closed issues:

• Unexpected memory accumulation when repeatedly running solve (#2147)

v6.71.0

OrdinaryDiffEq v6.71.0

Diff since v6.70.1

Merged pull requests:

• Fix out-of-place Rosenbrock23 (#2123) (@YingboMa)
• added methods ROS2PR, ROS2S, ROS3, ROS3PR, SCHOLZ4_7, ROS3PRL, ROS3PRL2 (#2126) (@cwittens)
• Bug in out-of-place version of Rodas3P: Update rosenbrock_perform_step.jl (#2128) (@gstein3m)
• Load SENSITIVITY_INTERP_MESSAGE from SciMLBase (#2136) (@LilithHafner)
• Improve callback modifier handling more (#2142) (@oscardssmith)

Closed issues:

• Improved traditional Rosenbrock–Wanner methods for stiff ODEs and DAEs by Joachim Rang (#2064)
• Rosenbrock solvers fail or interpolate poorly due to (algebraically) inconsistent u after callback (#2127)
• Rodas5P gives inaccurate results when integrating triangle wave (#2132)
• Rodas3P and Rodas23W are completely broken for out of place. (#2135)

v6.70.1

OrdinaryDiffEq v6.70.1

Diff since v6.70.0