v0.4.1

MadNLP v0.4.1

Diff since v0.4.0

Closed issues:

• Hessian information needed when using user-registered functions in JuMP (#115)
• MadNLP breaks Julia logging macros (#153)

Merged pull requests:

• add support to MOI 1.0 (#148) (@frapac)
• Add interface for custom scaling (#152) (@frapac)
• Do not overload Logging macros (#154) (@frapac)
• MadNLPPardisoMKL move to MadNLPPardiso.jl (#157) (@sshin23)
• ci version up (#158) (@sshin23)
• Unreduced + inertia-free fix (#159) (@sshin23)
• Bump MadNLP version to v0.4.1 (#160) (@sshin23)

v0.4.0

MadNLP v0.4.0

Diff since v0.3.0

Merged pull requests:

• Port MadNLP to MOI 0.10 and JuMP 0.22 (#117) (@frapac)
• add support to NLPModels 0.18 (#137) (@frapac)
• relax bound on Metis (#138) (@frapac)
• bump MadNLP version to 0.4.0 (#147) (@frapac)

v0.3.0

MadNLP v0.3.0

Diff since v0.2.0

Closed issues:

• Restructure dependencies in MadNLP (#32)
• Add support for dense nonlinear programming (#36)
• Retire NonlinearProgram in favor of NLPModels (#67)

Merged pull requests:

• Making linear solvers independent libraries (#40) (@sshin23)
• Add benchmarks (#45) (@sshin23)
• Compat update for NLPModels, SolverCore, and AmplNLReader (#48) (@sshin23)
• [MadNLPGPU] compat for CUDA updated (#49) (@sshin23)
• CompatHelper: bump compat for "StaticArrays" to "1" for package MadNLPMumps (#51) (@github-actions[bot])
• MadNLPHSL build with MKL (#54) (@sshin23)
• Add no_scaling option (#56) (@sshin23)
• Trigger github actions only on master and on open PRs (#57) (@frapac)
• Add AbstractKKTSystem structure (#58) (@frapac)
• Deactive logs in benchmark scripts (#59) (@frapac)
• benchmark improvement (#60) (@sshin23)
• barrier iterations (#61) (@sshin23)
• performance improvement (#64) (@sshin23)
• Add DenseKKTSystem structure (#65) (@frapac)
• benchmark-power fix (#68) (@sshin23)
• Retire NonlinearProgram (#69) (@sshin23)
• Allow to instantiate DenseKKTSystem on CUDA GPU (#73) (@frapac)
• timing error fix (#74) (@sshin23)
• rethrow bug fix (#75) (@sshin23)
• benchmark update (#76) (@sshin23)
• disabling rescale bound duals after penalty paramter update (#77) (@sshin23)
• change license to MIT (#79) (@sshin23)
• Check if varphi_trial, theta_trial is NaN (#80) (@sshin23)
• Use LapackCPU in MINLPTests (#81) (@sshin23)
• _get_fixed_variable_index update (#82) (@sshin23)
• Benchmark exclude explained (#84) (@sshin23)
• CUTEst benchmark tolerance 1e-8 (#87) (@sshin23)
• kkt error reduction (#88) (@sshin23)
• Add support for dense Cholesky (#89) (@sshin23)
• if_sufficient_progress fix for unconstrained problem (#90) (@sshin23)
• Add decode option in benchmark (#91) (@sshin23)
• Disable initial bound dual scaling (#92) (@sshin23)
• Filter acceptability function fix (#98) (@sshin23)
• [Breaking] Rename Solver as InteriorPointSolver (#99) (@frapac)
• Bump MadNLP version to v0.3.0 (#102) (@sshin23)
• [inertia] check if inertia is correct directly with KKTSystem (#104) (@frapac)

v0.2.0

MadNLP v0.2.0

Diff since v0.1.6

v0.1.6

MadNLP v0.1.6

Diff since v0.1.5

Closed issues:

• [CUDA] Add support for cuSOLVER's sytrs (#35)

Merged pull requests:

• CompatHelper: bump compat for "StaticArrays" to "1.0" (#16) (@github-actions[bot])
• CompatHelper: bump compat for "MPI" to "0.16" (#17) (@github-actions[bot])
• CompatHelper: bump compat for "IterativeSolvers" to "0.9" (#18) (@github-actions[bot])
• fixed plasmo objective value issue (#19) (@sshin23)
• schur bug fix (#20) (@sshin23)
• added resolving capability (#21) (@sshin23)
• user can give shared library path (#22) (@sshin23)
• Removing MPI and CUDA dependency (#23) (@sshin23)
• CompatHelper: bump compat for "NLPModels" to "0.14" (#24) (@github-actions[bot])
• CompatHelper: bump compat for "SolverTools" to "0.4" (#25) (@github-actions[bot])
• CompatHelper: bump compat for "SolverTools" to "0.5" (#27) (@github-actions[bot])
• CompatHelper: bump compat for "CUDA" to "3.0" (#28) (@github-actions[bot])
• Update for Plasmo v0.4. (#30) (@jalving)
• Update NLPModels and SolverCore (#31) (@frapac)
• Optimize copy of a SparseMatrixCOO to a dense matrix (#33) (@frapac)
• Fix MadNLP with IterativeSolvers 9.1 (#34) (@frapac)
• CompatHelper: bump compat for "NLPModels" to "0.15" (#37) (@github-actions[bot])
• Add support for cuSOLVER's sytrs (#38) (@sshin23)

v0.1.5

MadNLP v0.1.5

Diff since v0.1.4

v0.1.4

MadNLP v0.1.4

Diff since v0.1.3

Closed issues:

• Clash between Ipopt.jl and MadNLP.jl issued by MUMPS_seq_jll (#13)

Merged pull requests:

• Mumps version issue (#14) (@sshin23)

v0.1.3

MadNLP v0.1.3

Diff since v0.1.2

Merged pull requests:

• Replacing artifacts by jlls (#12) (@sshin23)

v0.1.2

MadNLP v0.1.2

Diff since v0.1.1

Merged pull requests:

• No-MKL support (#6) (@sshin23)
• Allow for CUDA 2.x (#7) (@michel2323)
• exclude farkas-related tests (#8) (@sshin23)
• test update (#9) (@sshin23)
• dense QR solver (CPU/GPU) (#10) (@sshin23)

v0.1.1

MadNLP v0.1.1

Diff since v0.1.0

Merged pull requests:

• Replace Travis with Github Actions (#4) (@michel2323)
• Replace Parameters.@with_kw by Base.@kwdef (#5) (@frapac)