v0.10.6

Krylov v0.10.6

Diff since v0.10.5

Merged pull requests:

• Add an implementation of USYMLQR (#886) (@amontoison)
• [documentation] Workspaces and solvers are now together (#1061) (@amontoison)
• [documentation] Add a reference for KrylovConstructor in the Krylov workspaces (#1062) (@amontoison)
• Improve interface.jl (#1063) (@amontoison)
• Add a function elapsed_allocation_time (#1064) (@amontoison)
• Fix the support of complex numbers in USYMLQR (#1065) (@amontoison)
• [USYMLQR] Remove the option transfer_to_usymcg (#1066) (@amontoison)
• [USYMLQR] Remove support for preconditioners (#1067) (@amontoison)
• Update a badge in README.md (#1069) (@amontoison)
• [documentation] Fix the docstring of USYMLQR (#1072) (@amontoison)
• Force inference of a type argument (#1074) (@timholy)
• Refactor allocate_if function to use type parameters (#1075) (@amontoison)
• Release 0.10.6 (#1077) (@amontoison)

Closed issues:

• USYMLQR (#164)
• Add a constant for the memory parameter in our metaprogramming script (#988)
• Another lnlq NaN case (#1054)
• Add a cross-reference for KrylovConstructor in the Krylov workspaces (#1059)

v0.10.5

Krylov v0.10.5

Diff since v0.10.4

Merged pull requests:

• Reset stats when reusing Krylov solver workspaces (#1058) (@farhadrclass)

v0.10.4

Krylov v0.10.4

Diff since v0.10.3

Merged pull requests:

• Allow vm and vn to be of different types (#1038) (@timholy)
• Improve inference for Double64 (#1045) (@timholy)
• Upgrade LNLQ (#1051) (@amontoison)
• fix typos (#1053) (@spaette)
• Update the workspaces to handle more storage types (#1056) (@amontoison)

Closed issues:

• KrylovConstructor: vm and vn should allow different types (#1037)
• Possible TRICG and TRIMR improvements (#1042)
• typos (#1048)
• lnlq can return NaNs (#1050)
• Unsupported keyword argument "memory" for GMRES and DQGMRES (#1055)

v0.10.3

Krylov v0.10.3

Diff since v0.10.2

Merged pull requests:

• [CI] Test Krylov.jl on GPU for Julia 1.12 (#1031) (@amontoison)
• [CI] Test Krylov.jl on Mac x86_64 (#1032) (@amontoison)
• Use kmul! and kldiv! for Krylov methods and processes (#1035) (@amontoison)
• Fix warm-start of TriCG and TriMR (#1040) (@timholy)

Closed issues:

• Add kmul! and kldiv! (#1030)

v0.10.2

Krylov v0.10.2

Diff since v0.10.1

Merged pull requests:

• Reuse the buffer for LAPACK routines on CPU (#940) (@amontoison)
• implementation Line Search with Negative Curvature Detection with direction for CR (#985) (@farhadrclass)
• implementation Line Search with Negative Curvature Detection with direction for CG (#1008) (@farhadrclass)
• implementation Line Search with Negative Curvature Detection with direction for MINRES (#1011) (@farhadrclass)
• Update the unit tests on Mac (#1016) (@amontoison)
• Update the link for the reference of bicgstab(l) (#1017) (@amontoison)
• [documentation] Update the example for the Poisson equation with halo regions (#1018) (@amontoison)
• [documentation] Additional modification in custom_workspaces.md (#1019) (@amontoison)
• Update make.jl (#1020) (@amontoison)
• Remove all allocations in the in-place version of block-MINRES (#1022) (@amontoison)
• Update ksizeof for workspaces of block-Krylov solvers (#1023) (@amontoison)
• Minres qlp npc (#1026) (@farhadrclass)

Closed issues:

• hermitian_lanczos and loss of orthogonality ? (#956)
• Is there a way to extract the Krylov Space? (#1007)
• Does Krylov use the S from LinearOperators? (#1009)
• How to Use Krylov.jl in Multithreading? (#1010)
• what is z in the docs? (#1012)
• GMRES using Float32 stagnates on Apple M2 (#1015)

v0.10.1

Krylov v0.10.1

Diff since v0.10.0

Merged pull requests:

• [documentation] Specify how to use krylov_workspace for each workspace (#1002) (@amontoison)
• [documentation] Tutorial on distributed Krylov solvers using MPI.jl (#1003) (@amontoison)
• [documentation] Update the routines for the support of MPI (#1004) (@amontoison)
• Add new methods for krylov_workspace and krylov_solve (#1005) (@amontoison)

v0.10.0

Krylov v0.10.0

Diff since v0.9.10

Breaking changes

• solve! is replaced by krylov_solve!
• Update the API for workspaces (the suffix Solver is now Workspace)
• Unify the API for the workspaces (memory and windows are always kwargs for the constructors)
• Ensure that S is a concrete type in the workspaces (KrylovWorkspace{T, FC, S})
• Update the API for workspace accessors (the functions are not exported anymore)
• Rename some workspace accessors (iteration_count, solution_count, Aprod_count and Atprod_count)

New features

• Add kdiv!, kscalcopy! and kdivcopy! for the support of custom workspaces
• Add a generic interface (krylov_workspace, krylov_solve and krylov_solve!)
• Add documentation on the generic API (see here)
• Add documentation on the workspace accessors (see here)
• Update the docstrings to add more cross-references
• Add a function Krylov.elapsed_time

Other changes

• Krylov.jl now requires Julia 1.10
• Update the reference of MinAres and CAr
• Return a nice error if we detect that the preconditioner is not SPD in CG

Merged pull requests:

• Implement Line Search with Negative Curvature Detection for MINRES Based on Liu et al. (2022) (#969) (@farhadrclass)
• Update the reference of MinAres (#971) (@amontoison)
• Update the reference of MinAres and CAr (#972) (@amontoison)
• Don't export Krylov.solve! (#973) (@amontoison)
• Require Julia 1.10 (#974) (@amontoison)
• Update the API for workspaces (#975) (@amontoison)
• S should be a concrete type in the workspaces (#977) (@amontoison)
• Update Breakage.yml for PRs from a fork (#978) (@amontoison)
• Add kdiv! -- kscalcopy! -- kdivcopy! (#980) (@amontoison)
• Add a generic interface for Krylov.jl (#983) (@amontoison)
• Rename the files krylov_solve.jl and test_solvers.jl (#989) (@amontoison)
• Use the suffix Workspace instead of Solver (#990) (@amontoison)
• Return a nice error if we detect the preconditioner is not SPD in CG (#991) (@amontoison)
• Add references to the generic interface in the docstrings (#992) (@amontoison)
• krylov_solvers -> krylov_workspaces (#993) (@amontoison)
• [documentation] Add more cross-references to the in-place methods (#994) (@amontoison)
• Remove aliases (#995) (@amontoison)
• Add a function krylov_elapsed_time (#996) (@amontoison)
• Add a new file krylov_show.jl (#997) (@amontoison)
• Update the API for workspace accessors (#998) (@amontoison)
• Polish the API for workspace accessors (#1000) (@amontoison)

Closed issues:

• Improve the documentation of functions Aprod and Atprod (#488)
• GMRES Fails Silently From Stagnation (#729)
• StaticArrays support (#766)
• Remove broadcast in Krylov methods (#930)
• Unexport Krylov.solve! with the next major release (#934)
• Add a function to get the timer directly from a KrylovWorkspace (#936)
• nshifts and S should be the last argument in the Krylov solvers (#946)
• Require Julia 1.10 for the next major release (#964)
• Bug Report: CR and CG Solvers Fail with Float16 Precision (#968)
• Update breakage.yml for PRs from a fork (#976)
• Add a generic constructor for KrylovSolver and BlockKrylovSolver (#979)
• Rename the files krylov_solve.jl and test_solvers.jl (#984)
• Rename the suffix of the workspaces *Solver -> *Workspace (#987)

v0.9.10

Krylov v0.9.10

Diff since v0.9.9

Merged pull requests:

• Use my new implementation of coytrito! in GPUArrays.jl (#950) (@amontoison)
• Use similar for dynamic allocations (#952) (@amontoison)
• [CI] Test Krylov.jl on Linux ARM (#953) (@amontoison)
• [documentation] Add an example with BlockArrays.jl (#954) (@amontoison)
• Add an option reorthogonalization in hermitian_lanczos (#957) (@amontoison)
• CR: Update iterate on first iteration for negative curvature with line search and when zero curvature is detected (#958) (@farhadrclass)
• Add more tests for the option warm-start (#961) (@amontoison)

Closed issues:

• Compatibility with BlockArrays (#605)
• [Benchmarks] Krylov.jl vs PETSc (#693)
• ComponentArrays support (#769)
• parallel gmres (#847)
• Add stationary methods (#904)
• scalar indexing error when using block_gmres on GPU (#913)
• How to solve Ax=b when A is a Sparse Array ? (#925)
• Add more advanced constructors for KrylovSolver (#933)
• Add a opnorm function (#942)
• [documentation] Add an example of preconditioned saddle point system with BlockArrays (#949)
• Use similar instead of S(undef, n) for dynamic allocations (#951)

v0.9.9

Krylov v0.9.9

Diff since v0.9.8

Merged pull requests:

• Add an implementation of BLOCK-MINRES (#884) (@amontoison)
• Matrix-free example example with a PDE (#911) (@amontoison)
• Update nvidia.jl (#916) (@amontoison)
• [documentation] Improve the docstring of SimpleStats (#917) (@amontoison)
• Improve the docstrings of Krylov.jl (#918) (@amontoison)
• [documentation] Add comments in krylov_solve.jl (#919) (@amontoison)
• [documentation] Add a link for the paper of BiCGstab(ℓ) (#920) (@amontoison)
• Update krylov_solve.jl (#921) (@amontoison)
• Fix a typo -- reflexion -> reflection (#922) (@amontoison)
• Use the infix operator to compute the timer (#923) (@amontoison)
• Update with JSOBestieTemplate for Krylov (#924) (@tmigot)
• Use rmul! instead of broadcast for kscal! (#926) (@amontoison)
• [documentation] Add a page about custom workspaces (#927) (@amontoison)
• Fix to_boundary for complex numbers (#928) (@amontoison)
• Remove broadcast in MINRES-QLP (#929) (@amontoison)
• Update trilqr.jl (#931) (@amontoison)
• Update with JSOBestieTemplate for Krylov (#938) (@tmigot)
• Upgrade TriCG, TriMR and GPMR (#943) (@amontoison)
• Upgrade the Krylov processes and detect breakdowns (#944) (@amontoison)
• Add a new constructor for all Krylov solvers (#947) (@amontoison)

Closed issues:

• Add a link for BICGSTAB(l) (#902)
• Support custom vector types (#905)
• Add comments in the source file krylov_solve.jl (#915)
• use of AbstractArray (#935)
• Shifted MINRES Allocations Issue (#937)
• Test CGLS with shifts on rectangular systems (#948)

v0.9.8

Krylov v0.9.8

Diff since v0.9.7

Merged pull requests:

• Remove old macros (#907) (@amontoison)
• Add a function knorm_elliptic (#909) (@amontoison)
• Improve the matrix-free example with FFT (#912) (@amontoison)
• Fix block_gmres on GPUs (#914) (@amontoison)

Closed issues:

• Remove the macros and directly use the functions starting with k... internally (#906)
• Add a function for sqrt(kdotr(n, v, z)) (#908)