Merge branch 'main' of github.com:SciML/LinearSolve.jl into ap/adjoint

.JuliaFormatter.toml

@@ -1,3 +1,4 @@
 style = "sciml"
 format_markdown = true
+format_docstrings = true
 annotate_untyped_fields_with_any = false

.github/dependabot.yml

@@ -7,4 +7,4 @@ updates:
       interval: "weekly"
     ignore:
       - dependency-name: "crate-ci/typos"
-        update-types: ["version-update:semver-patch"]
+        update-types: ["version-update:semver-patch", "version-update:semver-minor"]

.github/workflows/CI.yml

@@ -32,7 +32,7 @@ jobs:
       - uses: julia-actions/setup-julia@v1
         with:
           version: ${{ matrix.version }}
-      - uses: actions/cache@v3
+      - uses: actions/cache@v4
         env:
           cache-name: cache-artifacts
         with:

.github/workflows/Downgrade.yml

@@ -0,0 +1,31 @@
+name: Downgrade
+on:
+  pull_request:
+    branches:
+      - main
+    paths-ignore:
+      - 'docs/**'
+  push:
+    branches:
+      - main
+    paths-ignore:
+      - 'docs/**'
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        version: ['1']
+        group:
+          - Core
+    steps:
+      - uses: actions/checkout@v4
+      - uses: julia-actions/setup-julia@v1
+        with:
+          version: ${{ matrix.version }}
+      - uses: julia-actions/julia-downgrade-compat@v1
+#        if: ${{ matrix.version == '1.6' }}
+        with:
+          skip: Pkg,TOML
+      - uses: julia-actions/julia-buildpkg@v1
+      - uses: julia-actions/julia-runtest@v1

.github/workflows/Downstream.yml

@@ -12,6 +12,7 @@ jobs:
     env:
       GROUP: ${{ matrix.package.group }}
     strategy:
+      fail-fast: false
       matrix:
         julia-version: [1]
         os: [ubuntu-latest]
@@ -20,7 +21,7 @@ jobs:
           - {user: SciML, repo: ModelingToolkit.jl, group: All}
           - {user: SciML, repo: SciMLSensitivity.jl, group: Core1}
           - {user: SciML, repo: BoundaryValueDiffEq.jl, group: All}
-
+          - {user: SciML, repo: NonlinearSolve.jl, group: All}
     steps:
       - uses: actions/checkout@v4
       - uses: julia-actions/setup-julia@v1
@@ -50,6 +51,9 @@ jobs:
             @info "Not compatible with this release. No problem." exception=err
             exit(0)  # Exit immediately, as a success
           end
+        env:
+          RETESTITEMS_NWORKERS: 4
+          RETESTITEMS_NWORKER_THREADS: 2
       - uses: julia-actions/julia-processcoverage@v1
       - uses: codecov/codecov-action@v3
         with:

.github/workflows/SpellCheck.yml

@@ -10,4 +10,4 @@ jobs:
       - name: Checkout Actions Repository
         uses: actions/checkout@v4
       - name: Check spelling
-        uses: crate-ci/typos@v1.16.23 
+        uses: crate-ci/typos@v1.18.0 

Project.toml

@@ -1,7 +1,7 @@
 name = "LinearSolve"
 uuid = "7ed4a6bd-45f5-4d41-b270-4a48e9bafcae"
 authors = ["SciML"]
-version = "2.22.0"
+version = "2.25.0"
 
 [deps]
 ArrayInterface = "4fba245c-0d91-5ea0-9b3e-6abc04ee57a9"
@@ -22,7 +22,6 @@ PrecompileTools = "aea7be01-6a6a-4083-8856-8a6e6704d82a"
 Preferences = "21216c6a-2e73-6563-6e65-726566657250"
 RecursiveFactorization = "f2c3362d-daeb-58d1-803e-2bc74f2840b4"
 Reexport = "189a3867-3050-52da-a836-e630ba90ab69"
-Requires = "ae029012-a4dd-5104-9daa-d747884805df"
 SciMLBase = "0bca4576-84f4-4d90-8ffe-ffa030f20462"
 SciMLOperators = "c0aeaf25-5076-4817-a8d5-81caf7dfa961"
 Setfield = "efcf1570-3423-57d1-acb7-fd33fddbac46"
@@ -63,53 +62,53 @@ LinearSolveRecursiveArrayToolsExt = "RecursiveArrayTools"
 [compat]
 AllocCheck = "0.1"
 Aqua = "0.8"
-ArrayInterface = "7.4.11"
-BandedMatrices = "1"
-BlockDiagonals = "0.1"
+ArrayInterface = "7.7"
+BandedMatrices = "1.5"
+BlockDiagonals = "0.1.42"
 CUDA = "5"
-ConcreteStructs = "0.2"
-DocStringExtensions = "0.9"
-EnumX = "1"
-Enzyme = "0.11"
-EnzymeCore = "0.6"
+ConcreteStructs = "0.2.3"
+DocStringExtensions = "0.9.3"
+EnumX = "1.0.4"
+Enzyme = "0.11.15"
+EnzymeCore = "0.6.5"
 FastAlmostBandedMatrices = "0.1"
 FastLapackInterface = "2"
-FiniteDiff = "2"
-ForwardDiff = "0.10"
-GPUArraysCore = "0.1"
+FiniteDiff = "2.22"
+ForwardDiff = "0.10.36"
+GPUArraysCore = "0.1.6"
 HYPRE = "1.4.0"
-InteractiveUtils = "1.6"
+InteractiveUtils = "1.10"
 IterativeSolvers = "0.9.3"
-JET = "0.8"
-KLU = "0.3.0, 0.4"
-KernelAbstractions = "0.9"
+JET = "0.8.28"
+KLU = "0.5"
+KernelAbstractions = "0.9.16"
 Krylov = "0.9"
 KrylovKit = "0.6"
-Libdl = "1.6"
-LinearAlgebra = "1.9"
+Libdl = "1.10"
+LinearAlgebra = "1.10"
 MPI = "0.20"
 Metal = "0.5"
 MultiFloats = "1"
 Pardiso = "0.5"
 Pkg = "1"
-PrecompileTools = "1"
-Preferences = "1"
+PrecompileTools = "1.2"
+Preferences = "1.4"
 Random = "1"
-RecursiveArrayTools = "2, 3"
-RecursiveFactorization = "0.2.8"
+RecursiveArrayTools = "3.8"
+RecursiveFactorization = "0.2.14"
 Reexport = "1"
-Requires = "1"
 SafeTestsets = "0.1"
-SciMLBase = "2"
-SciMLOperators = "0.3"
+SciMLBase = "2.23.0"
+SciMLOperators = "0.3.7"
 Setfield = "1"
-SparseArrays = "1.9"
+SparseArrays = "1.10"
 Sparspak = "0.3.6"
-StaticArrays = "1"
-StaticArraysCore = "1"
+StableRNGs = "1"
+StaticArrays = "1.5"
+StaticArraysCore = "1.4.2"
 Test = "1"
 UnPack = "1"
-julia = "1.9"
+julia = "1.10"
 
 [extras]
 AllocCheck = "9b6a8646-10ed-4001-bbdc-1d2f46dfbb1a"
@@ -133,8 +132,9 @@ Pkg = "44cfe95a-1eb2-52ea-b672-e2afdf69b78f"
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
 RecursiveArrayTools = "731186ca-8d62-57ce-b412-fbd966d074cd"
 SafeTestsets = "1bc83da4-3b8d-516f-aca4-4fe02f6d838f"
+StableRNGs = "860ef19b-820b-49d6-a774-d7a799459cd3"
 StaticArrays = "90137ffa-7385-5640-81b9-e52037218182"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 
 [targets]
-test = ["Aqua", "Test", "IterativeSolvers", "InteractiveUtils", "JET", "KrylovKit", "Pkg", "Random", "SafeTestsets", "MultiFloats", "ForwardDiff", "HYPRE", "MPI", "BlockDiagonals", "Enzyme", "FiniteDiff", "BandedMatrices", "FastAlmostBandedMatrices", "StaticArrays", "AllocCheck"]
+test = ["Aqua", "Test", "IterativeSolvers", "InteractiveUtils", "JET", "KrylovKit", "Pkg", "Random", "SafeTestsets", "MultiFloats", "ForwardDiff", "HYPRE", "MPI", "BlockDiagonals", "Enzyme", "FiniteDiff", "BandedMatrices", "FastAlmostBandedMatrices", "StaticArrays", "AllocCheck", "StableRNGs"]

benchmarks/applelu.jl

@@ -23,7 +23,7 @@ algs = [
     GenericLUFactorization(),
     RFLUFactorization(),
     AppleAccelerateLUFactorization(),
-    MetalLUFactorization(),
+    MetalLUFactorization()
 ]
 res = [Float32[] for i in 1:length(algs)]
 

benchmarks/lu.jl

@@ -24,7 +24,7 @@ algs = [
     RFLUFactorization(),
     MKLLUFactorization(),
     FastLUFactorization(),
-    SimpleLUFactorization(),
+    SimpleLUFactorization()
 ]
 res = [Float64[] for i in 1:length(algs)]
 

benchmarks/sparselu.jl

@@ -36,7 +36,7 @@ algs = [
     UMFPACKFactorization(),
     KLUFactorization(),
     MKLPardisoFactorize(),
-    SparspakFactorization(),
+    SparspakFactorization()
 ]
 cols = [:red, :blue, :green, :magenta, :turqoise] # one color per alg
 lst = [:dash, :solid, :dashdot] # one line style per dim
@@ -65,7 +65,8 @@ function run_and_plot(; dims = [1, 2, 3], kmax = 12)
             u0 = rand(rng, n)
 
             for j in 1:length(algs)
-                bt = @belapsed solve(prob, $(algs[j])).u setup=(prob = LinearProblem(copy($A),
+                bt = @belapsed solve(prob, $(algs[j])).u setup=(prob = LinearProblem(
+                    copy($A),
                     copy($b);
                     u0 = copy($u0),
                     alias_A = true,

docs/pages.jl

@@ -2,14 +2,14 @@
 
 pages = ["index.md",
     "Tutorials" => Any["tutorials/linear.md"
-        "tutorials/caching_interface.md"],
+                       "tutorials/caching_interface.md"],
     "Basics" => Any["basics/LinearProblem.md",
         "basics/common_solver_opts.md",
         "basics/OperatorAssumptions.md",
         "basics/Preconditioners.md",
         "basics/FAQ.md"],
     "Solvers" => Any["solvers/solvers.md"],
     "Advanced" => Any["advanced/developing.md"
-        "advanced/custom.md"],
-    "Release Notes" => "release_notes.md",
+                      "advanced/custom.md"],
+    "Release Notes" => "release_notes.md"
 ]

docs/src/advanced/developing.md

@@ -18,7 +18,7 @@ basic machinery. A simplified version is:
 struct MyLUFactorization{P} <: SciMLBase.AbstractLinearAlgorithm end
 
 function init_cacheval(alg::MyLUFactorization, A, b, u, Pl, Pr, maxiters, abstol, reltol,
-    verbose)
+        verbose)
     lu!(convert(AbstractMatrix, A))
 end
 

docs/src/basics/Preconditioners.md

@@ -83,13 +83,13 @@ The following preconditioners match the interface of LinearSolve.jl.
   - [Preconditioners.CholeskyPreconditioner(A, i)](https://github.com/JuliaLinearAlgebra/Preconditioners.jl):
     An incomplete Cholesky preconditioner with cut-off level `i`. Requires `A` as
     a `AbstractMatrix` and positive semi-definite.
-  - [AlgebraicMultiGrid](https://github.com/JuliaLinearAlgebra/AlgebraicMultigrid.jl):
+  - [AlgebraicMultigrid](https://github.com/JuliaLinearAlgebra/AlgebraicMultigrid.jl):
     Implementations of the algebraic multigrid method. Must be converted to a
-    preconditioner via `AlgebraicMultiGrid.aspreconditioner(AlgebraicMultiGrid.precmethod(A))`.
+    preconditioner via `AlgebraicMultigrid.aspreconditioner(AlgebraicMultigrid.precmethod(A))`.
     Requires `A` as a `AbstractMatrix`. Provides the following methods:
 
-      + `AlgebraicMultiGrid.ruge_stuben(A)`
-      + `AlgebraicMultiGrid.smoothed_aggregation(A)`
+      + `AlgebraicMultigrid.ruge_stuben(A)`
+      + `AlgebraicMultigrid.smoothed_aggregation(A)`
   - [PyAMG](https://github.com/cortner/PyAMG.jl):
     Implementations of the algebraic multigrid method. Must be converted to a
     preconditioner via `PyAMG.aspreconditioner(PyAMG.precmethod(A))`.
@@ -111,3 +111,9 @@ The following preconditioners match the interface of LinearSolve.jl.
     preconditioners which supports distributed computing via MPI. These can be
     written using the LinearSolve.jl interface choosing algorithms like `HYPRE.ILU`
     and `HYPRE.BoomerAMG`.
+  - [KrylovPreconditioners.jl](https://github.com/JuliaSmoothOptimizers/KrylovPreconditioners.jl/): Provides GPU-ready
+    preconditioners via KernelAbstractions.jl. At the time of writing the package provides the following methods:
+
+      + Incomplete Cholesky decomposition `KrylovPreconditioners.kp_ic0(A)`
+      + Incomplete LU decomposition `KrylovPreconditioners.kp_ilu0(A)`
+      + Block Jacobi `KrylovPreconditioners.BlockJacobiPreconditioner(A, nblocks, device)`

ext/LinearSolveBandedMatricesExt.jl

@@ -2,7 +2,8 @@ module LinearSolveBandedMatricesExt
 
 using BandedMatrices, LinearAlgebra, LinearSolve
 import LinearSolve: defaultalg,
-    do_factorization, init_cacheval, DefaultLinearSolver, DefaultAlgorithmChoice
+                    do_factorization, init_cacheval, DefaultLinearSolver,
+                    DefaultAlgorithmChoice
 
 # Defaults for BandedMatrices
 function defaultalg(A::BandedMatrix, b, oa::OperatorAssumptions{Bool})
@@ -35,14 +36,14 @@ for alg in (:SVDFactorization, :MKLLUFactorization, :DiagonalFactorization,
     :AppleAccelerateLUFactorization, :CholeskyFactorization)
     @eval begin
         function init_cacheval(::$(alg), ::BandedMatrix, b, u, Pl, Pr, maxiters::Int,
-            abstol, reltol, verbose::Bool, assumptions::OperatorAssumptions)
+                abstol, reltol, verbose::Bool, assumptions::OperatorAssumptions)
             return nothing
         end
     end
 end
 
 function init_cacheval(::LUFactorization, A::BandedMatrix, b, u, Pl, Pr, maxiters::Int,
-    abstol, reltol, verbose::Bool, assumptions::OperatorAssumptions)
+        abstol, reltol, verbose::Bool, assumptions::OperatorAssumptions)
     return lu(similar(A, 0, 0))
 end
 
@@ -54,8 +55,8 @@ for alg in (:SVDFactorization, :MKLLUFactorization, :DiagonalFactorization,
     :AppleAccelerateLUFactorization, :QRFactorization, :LUFactorization)
     @eval begin
         function init_cacheval(::$(alg), ::Symmetric{<:Number, <:BandedMatrix}, b, u, Pl,
-            Pr, maxiters::Int, abstol, reltol, verbose::Bool,
-            assumptions::OperatorAssumptions)
+                Pr, maxiters::Int, abstol, reltol, verbose::Bool,
+                assumptions::OperatorAssumptions)
             return nothing
         end
     end

ext/LinearSolveBlockDiagonalsExt.jl

@@ -3,7 +3,7 @@ module LinearSolveBlockDiagonalsExt
 using LinearSolve, BlockDiagonals
 
 function LinearSolve.init_cacheval(alg::SimpleGMRES{false}, A::BlockDiagonal, b, args...;
-    kwargs...)
+        kwargs...)
     @assert ndims(A)==2 "ndims(A) == $(ndims(A)). `A` must have ndims == 2."
     # We need to perform this check even when `zeroinit == true`, since the type of the
     # cache is dependent on whether we are able to use the specialized dispatch.

ext/LinearSolveCUDAExt.jl

@@ -6,7 +6,7 @@ using LinearSolve.LinearAlgebra, LinearSolve.SciMLBase, LinearSolve.ArrayInterfa
 using SciMLBase: AbstractSciMLOperator
 
 function SciMLBase.solve!(cache::LinearSolve.LinearCache, alg::CudaOffloadFactorization;
-    kwargs...)
+        kwargs...)
     if cache.isfresh
         fact = qr(CUDA.CuArray(cache.A))
         cache.cacheval = fact
@@ -18,8 +18,8 @@ function SciMLBase.solve!(cache::LinearSolve.LinearCache, alg::CudaOffloadFactor
 end
 
 function LinearSolve.init_cacheval(alg::CudaOffloadFactorization, A, b, u, Pl, Pr,
-    maxiters::Int, abstol, reltol, verbose::Bool,
-    assumptions::OperatorAssumptions)
+        maxiters::Int, abstol, reltol, verbose::Bool,
+        assumptions::OperatorAssumptions)
     qr(CUDA.CuArray(A))
 end