v0.7 upgrade, comments, bugfixes and cleanup

.travis.yml

@@ -4,6 +4,8 @@ os:
   - linux
 julia:
   - 0.6
+  - 0.7
+  - 1.0
   - nightly
 notifications:
   email: false
@@ -14,7 +16,7 @@ git:
 ## (tests will run but not make your overall status red)
 matrix:
   allow_failures:
-  - julia: nightly
+    - julia: nightly
 
 ## uncomment and modify the following lines to manually install system packages
 addons:
@@ -25,10 +27,10 @@ addons:
 #  - if [ $TRAVIS_OS_NAME = osx ]; then brew install gcc; fi
 
 ## uncomment the following lines to override the default test script
-#script:
-#  - julia -e 'Pkg.clone(pwd()); Pkg.build("AMG"); Pkg.test("AMG"; coverage=true)'
+script:
+  - julia -e 'VERSION >= v"0.7-" && using Pkg; Pkg.clone(pwd()); Pkg.build("AlgebraicMultigrid"); Pkg.test("AlgebraicMultigrid"; coverage=true)'
 after_success:
   # push coverage results to Coveralls
-  - julia -e 'cd(Pkg.dir("AMG")); Pkg.add("Coverage"); using Coverage; Coveralls.submit(Coveralls.process_folder())'
+  - julia -e 'cd(Pkg.dir("AlgebraicMultigrid")); Pkg.add("Coverage"); using Coverage; Coveralls.submit(Coveralls.process_folder())'
   # push coverage results to Codecov
-  - julia -e 'cd(Pkg.dir("AMG")); Pkg.add("Coverage"); using Coverage; Codecov.submit(Codecov.process_folder())'
+  - julia -e 'cd(Pkg.dir("AlgebraicMultigrid")); Pkg.add("Coverage"); using Coverage; Codecov.submit(Codecov.process_folder())'

REQUIRE

@@ -1,2 +1,3 @@
 julia 0.6
-IterativeSolvers 0.4.1
+IterativeSolvers 0.6.0
+Compat 1.0.0

src/AMG.jl → src/AlgebraicMultigrid.jl

@@ -1,25 +1,34 @@
-module AMG
+module AlgebraicMultigrid
 
 import IterativeSolvers: gauss_seidel!
+using Compat, Compat.LinearAlgebra
+using Compat.SparseArrays, Compat.Printf
 using Base.Threads
 
+using Compat: rmul!
+
+if VERSION < v"0.7-"
+    const mul! = A_mul_B!
+end
+
 const MT = false
+const AMG = AlgebraicMultigrid
 
 include("utils.jl")
 export approximate_spectral_radius
 
 include("strength.jl")
-export strength_of_connection, Classical, SymmetricStrength
+export Classical, SymmetricStrength
 
 include("splitting.jl")
-export split_nodes, RS
+export RS
 
 include("gallery.jl")
 export poisson
 
 include("smoother.jl")
 export GaussSeidel, SymmetricSweep, ForwardSweep, BackwardSweep,
-        smooth_prolongator, JacobiProlongation
+        JacobiProlongation
 
 include("multilevel.jl")
 export solve
@@ -28,7 +37,7 @@ include("classical.jl")
 export ruge_stuben
 
 include("aggregate.jl")
-export aggregation, StandardAggregation
+export StandardAggregation
 
 include("aggregation.jl")
 export fit_candidates, smoothed_aggregation

src/aggregate.jl

@@ -1,7 +1,7 @@
 struct StandardAggregation
 end
 
-function aggregation(::StandardAggregation, S::SparseMatrixCSC{T,R}) where {T,R}
+function (::StandardAggregation)(S::SparseMatrixCSC{T,R}) where {T,R}
 
     n = size(S, 1)
     x = zeros(R, n)
@@ -97,7 +97,7 @@ function aggregation(::StandardAggregation, S::SparseMatrixCSC{T,R}) where {T,R}
     if minimum(x) == -1
         mask = x .!= -1
         I = collect(R, 1:n)[mask]
-        J = x[mask] + R(1)
+        J = x[mask] .+ R(1)
         #J = x[mask] + 1
         V = ones(eltype(S), length(J))
         AggOp = sparse(J,I,V,N,M)

src/aggregation.jl

@@ -1,17 +1,16 @@
-function smoothed_aggregation(A::SparseMatrixCSC{T,V},
+function smoothed_aggregation(A::SparseMatrixCSC{T,V}, ::Type{Val{bs}}=Val{1},
                         symmetry = HermitianSymmetry(),
                         strength = SymmetricStrength(),
                         aggregate = StandardAggregation(),
                         smooth = JacobiProlongation(4.0/3.0),
                         presmoother = GaussSeidel(),
                         postsmoother = GaussSeidel(),
-                        improve_candidates = GaussSeidel(4),
+                        improve_candidates = GaussSeidel(iter=4),
                         max_levels = 10,
                         max_coarse = 10,
                         diagonal_dominance = false,
                         keep = false,
-                        coarse_solver = Pinv()) where {T,V}
-
+                        coarse_solver = Pinv) where {T,V,bs}
 
     n = size(A, 1)
     # B = kron(ones(n, 1), eye(1))
@@ -30,19 +29,23 @@ function smoothed_aggregation(A::SparseMatrixCSC{T,V},
 
     levels = Vector{Level{T,V}}()
     bsr_flag = false
+    w = MultiLevelWorkspace(Val{bs}, eltype(A))
 
     while length(levels) + 1 < max_levels && size(A, 1) > max_coarse
+        residual!(w, size(A, 1))
         A, B, bsr_flag = extend_hierarchy!(levels, strength, aggregate, smooth,
                                 improve_candidates, diagonal_dominance,
                                 keep, A, B, symmetry, bsr_flag)
+        coarse_x!(w, size(A, 1))
+        coarse_b!(w, size(A, 1))
         #=if size(A, 1) <= max_coarse
             break
         end=#
     end
     #=A, B = extend_hierarchy!(levels, strength, aggregate, smooth,
                             improve_candidates, diagonal_dominance,
                             keep, A, B, symmetry)=#
-    MultiLevel(levels, A, presmoother, postsmoother)
+    MultiLevel(levels, A, coarse_solver(A), presmoother, postsmoother, w)
 end
 
 struct HermitianSymmetry
@@ -54,18 +57,18 @@ function extend_hierarchy!(levels, strength, aggregate, smooth,
                             symmetry, bsr_flag)
 
     # Calculate strength of connection matrix
-    S = strength_of_connection(strength, A, bsr_flag)
+    S = strength(A, bsr_flag)
 
     # Aggregation operator
-    AggOp = aggregation(aggregate, S)
+    AggOp = aggregate(S)
     # b = zeros(eltype(A), size(A, 1))
 
     # Improve candidates
     b = zeros(size(A,1))
-    relax!(improve_candidates, A, B, b)
+    improve_candidates(A, B, b)
     T, B = fit_candidates(AggOp, B)
 
-    P = smooth_prolongator(smooth, A, T, S, B)
+    P = smooth(A, T, S, B)
     R = construct_R(symmetry, P)
     push!(levels, Level(A, P, R))
 
@@ -77,11 +80,11 @@ function extend_hierarchy!(levels, strength, aggregate, smooth,
 
     A, B, bsr_flag
 end
-construct_R(::HermitianSymmetry, P) = P'
+construct_R(::HermitianSymmetry, P) = copy(P')
 
 function fit_candidates(AggOp, B, tol = 1e-10)
 
-    A = AggOp.'
+    A = copy(AggOp')
     n_fine, n_coarse = size(A)
     n_col = n_coarse