Remove Static.jl and StaticArrayInterface; use FastBroadcast Serial/Threaded

Replace Static.True()/False() threading dispatch with FastBroadcast's
Serial/Threaded types across DiffEqBase, OrdinaryDiffEqCore, and all
subpackages. Remove StaticArrayInterface from Core and Differentiation.
Require FastBroadcast >= 1.1.

Co-Authored-By: Chris Rackauckas <accounts@chrisrackauckas.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: ChrisRackauckas

lib/DelayDiffEq/Project.toml

@@ -88,7 +88,7 @@ OrdinaryDiffEqSSPRK = "1.5.0"
 OrdinaryDiffEqTsit5 = "1.4.0"
 OrdinaryDiffEqVerner = "1.5.0"
 Random = "<0.0.1, 1"
-RecursiveArrayTools = "3.52"
+RecursiveArrayTools = "3"
 Reexport = "1.0"
 SafeTestsets = "0.1.0"
 SciMLBase = "2.115.0"
@@ -109,10 +109,10 @@ julia = "1.10"
 ADTypes = "47edcb42-4c32-4615-8424-f2b9edc5f35b"
 DiffEqNoiseProcess = "77a26b50-5914-5dd7-bc55-306e6241c503"
 StochasticDiffEqCore = "19c5a474-6cd1-4a5f-be79-46dc34e54d7f"
-StochasticDiffEqHighOrder = "0520c28c-50fd-4d16-9c96-902fc80b3bab"
-StochasticDiffEqImplicit = "5080b986-4c76-4669-b5dc-373a41579d5b"
+StochasticDiffEqHighOrder = "2396deb8-7031-47e3-b5f0-b712c0a6b4cc"
+StochasticDiffEqImplicit = "b4b012ad-4b5e-4e99-a72e-7e09e7e0b800"
 StochasticDiffEqLowOrder = "d15fe365-ce7f-450a-828a-7985bd5b681b"
-StochasticDiffEqROCK = "db241ea8-0e6b-4abc-8f2d-1adff2294fd9"
+StochasticDiffEqROCK = "c3d49e4e-9e6b-4ee6-becc-4dd72c61f2cc"
 SparseArrays = "2f01184e-e22b-5df5-ae63-d93ebab69eaf"
 Aqua = "4c88cf16-eb10-579e-8560-4a9242c79595"
 DDEProblemLibrary = "f42792ee-6ffc-4e2a-ae83-8ee2f22de800"

lib/DiffEqBase/Project.toml

@@ -26,7 +26,6 @@ SciMLLogging = "a6db7da4-7206-11f0-1eab-35f2a5dbe1d1"
 SciMLOperators = "c0aeaf25-5076-4817-a8d5-81caf7dfa961"
 SciMLStructures = "53ae85a6-f571-4167-b2af-e1d143709226"
 Setfield = "efcf1570-3423-57d1-acb7-fd33fddbac46"
-Static = "aedffcd0-7271-4cad-89d0-dc628f76c6d3"
 StaticArraysCore = "1e83bf80-4336-4d27-bf5d-d5a4f845583c"
 SymbolicIndexingInterface = "2efcf032-c050-4f8e-a9bb-153293bab1f5"
 TruncatedStacktraces = "781d530d-4396-4725-bb49-402e4bee1e77"
@@ -89,7 +88,7 @@ FastPower = "1.1"
 FlexUnits = "0.4"
 ForwardDiff = "0.10, 1"
 FunctionWrappers = "1.0"
-FunctionWrappersWrappers = "1"
+FunctionWrappersWrappers = "0.1, 1"
 GTPSA = "1.4"
 GeneralizedGenerated = "0.3"
 InteractiveUtils = "1.9"
@@ -106,7 +105,7 @@ Pkg = "1.9"
 PrecompileTools = "1"
 Printf = "1.9"
 Random = "1.9"
-RecursiveArrayTools = "3.52"
+RecursiveArrayTools = "3.1"
 Reexport = "1.0"
 ReverseDiff = "1"
 SafeTestsets = "0.1"
@@ -116,7 +115,6 @@ SciMLOperators = "1"
 SciMLStructures = "1.5"
 Setfield = "1"
 SparseArrays = "1.9"
-Static = "1"
 StaticArrays = "1"
 StaticArraysCore = "1.4"
 SymbolicIndexingInterface = "0.3.39"

lib/DiffEqBase/ext/DiffEqBaseForwardDiffExt.jl

@@ -4,7 +4,7 @@ using DiffEqBase, ForwardDiff
 using DiffEqBase.ArrayInterface
 using DiffEqBase: Void, FunctionWrappersWrappers, OrdinaryDiffEqTag,
     AbstractTimeseriesSolution,
-    RecursiveArrayTools, reduce_tup, _promote_tspan, has_continuous_callback
+    RecursiveArrayTools, _promote_tspan, has_continuous_callback
 import DiffEqBase: hasdualpromote, wrapfun_oop, wrapfun_iip, prob2dtmin,
     promote_tspan, ODE_DEFAULT_NORM
 import SciMLBase: isdualtype, DualEltypeChecker, sse, __sum

lib/DiffEqBase/src/DiffEqBase.jl

@@ -21,10 +21,7 @@ using FunctionWrappers: FunctionWrapper
 using MuladdMacro
 
 
-using FastBroadcast: @..
-using Static: True, False
-
-using Static: reduce_tup
+using FastBroadcast: @.., Serial, Threaded
 
 import RecursiveArrayTools
 import TruncatedStacktraces

lib/DiffEqBase/src/calculate_residuals.jl

@@ -90,7 +90,7 @@ end
 # Inplace Versions
 
 """
-    DiffEqBase.calculate_residuals!(out, ũ, u₀, u₁, α, ρ, thread=False())
+    DiffEqBase.calculate_residuals!(out, ũ, u₀, u₁, α, ρ, thread=Serial())
 
 Save element-wise residuals
 ```math
@@ -99,13 +99,13 @@ Save element-wise residuals
 in `out`.
 
 The argument `thread` determines whether internal broadcasting on
-appropriate CPU arrays should be serial (`thread = False()`, default)
-or use multiple threads (`thread = True()`) when Julia is started
+appropriate CPU arrays should be serial (`thread = Serial()`, default)
+or use multiple threads (`thread = Threaded()`) when Julia is started
 with multiple threads.
 """
 @inline function calculate_residuals!(
         out, ũ, u₀, u₁, α, ρ, internalnorm, t,
-        thread::Union{False, True} = False()
+        thread = Serial()
     )
     @.. broadcast = false thread = thread out = calculate_residuals(
         ũ, u₀, u₁, α, ρ, internalnorm,
@@ -114,18 +114,8 @@ with multiple threads.
     return nothing
 end
 
-@inline function calculate_residuals!(
-        out::Array, ũ::Array, u₀::Array, u₁::Array, α::Number,
-        ρ::Number, internalnorm::F, t, ::False
-    ) where {F}
-    @inbounds @simd ivdep for i in eachindex(out, ũ, u₀, u₁)
-        out[i] = calculate_residuals(ũ[i], u₀[i], u₁[i], α, ρ, internalnorm, t)
-    end
-    return nothing
-end
-
 """
-    calculate_residuals!(out, u₀, u₁, α, ρ, thread=False())
+    calculate_residuals!(out, u₀, u₁, α, ρ, thread=Serial())
 
 Save element-wise residuals
 ```math
@@ -134,34 +124,34 @@ Save element-wise residuals
 in `out`.
 
 The argument `thread` determines whether internal broadcasting on
-appropriate CPU arrays should be serial (`thread = False()`, default)
-or use multiple threads (`thread = True()`) when Julia is started
+appropriate CPU arrays should be serial (`thread = Serial()`, default)
+or use multiple threads (`thread = Threaded()`) when Julia is started
 with multiple threads.
 """
 @inline function calculate_residuals!(
         out, u₀, u₁, α, ρ, internalnorm, t,
-        thread::Union{False, True} = False()
+        thread = Serial()
     )
     @.. broadcast = false thread = thread out = calculate_residuals(u₀, u₁, α, ρ, internalnorm, t)
     return nothing
 end
 
 """
-    calculate_residuals!(out, E₁, E₂, u₀, u₁, α, ρ, δ, scalarnorm, thread=False())
+    calculate_residuals!(out, E₁, E₂, u₀, u₁, α, ρ, δ, scalarnorm, thread=Serial())
 
 Calculate element-wise residuals
 ```math
 \\frac{δ E₁ + E₂}{α+\\max{scalarnorm(u₀),scalarnorm(u₁)}*ρ}.
 ```
 
 The argument `thread` determines whether internal broadcasting on
-appropriate CPU arrays should be serial (`thread = False()`, default)
-or use multiple threads (`thread = True()`) when Julia is started
+appropriate CPU arrays should be serial (`thread = Serial()`, default)
+or use multiple threads (`thread = Threaded()`) when Julia is started
 with multiple threads.
 """
 @inline function calculate_residuals!(
         out, E₁, E₂, u₀, u₁, α, ρ, δ, scalarnorm, t,
-        thread::Union{False, True} = False()
+        thread = Serial()
     )
     @.. broadcast = false thread = thread out = calculate_residuals(
         E₁, E₂, u₀, u₁, α, ρ, δ,

lib/OrdinaryDiffEqAdamsBashforthMoulton/Project.toml

@@ -6,7 +6,6 @@ version = "1.11.0"
 [deps]
 MuladdMacro = "46d2c3a1-f734-5fdb-9937-b9b9aeba4221"
 Polyester = "f517fe37-dbe3-4b94-8317-1923a5111588"
-Static = "aedffcd0-7271-4cad-89d0-dc628f76c6d3"
 FastBroadcast = "7034ab61-46d4-4ed7-9d0f-46aef9175898"
 OrdinaryDiffEqLowOrderRK = "1344f307-1e59-4825-a18e-ace9aa3fa4c6"
 RecursiveArrayTools = "731186ca-8d62-57ce-b412-fbd966d074cd"
@@ -36,7 +35,6 @@ MuladdMacro = "0.2"
 Polyester = "0.7"
 SciMLBase = "2.116"
 OrdinaryDiffEqCore = "3"
-Static = "1.2"
 OrdinaryDiffEqLowOrderRK = "1.5.0"
 julia = "1.10"
 RecursiveArrayTools = "3.52"

lib/OrdinaryDiffEqAdamsBashforthMoulton/src/OrdinaryDiffEqAdamsBashforthMoulton.jl

@@ -14,7 +14,7 @@ import OrdinaryDiffEqCore: OrdinaryDiffEqMutableCache, OrdinaryDiffEqConstantCac
 import OrdinaryDiffEqLowOrderRK: BS3ConstantCache, BS3Cache, RK4ConstantCache, RK4Cache
 import RecursiveArrayTools: recursivefill!
 using MuladdMacro, FastBroadcast
-import Static: False
+using FastBroadcast: Serial
 import OrdinaryDiffEqCore
 
 using Reexport

lib/OrdinaryDiffEqAdamsBashforthMoulton/src/adams_bashforth_moulton_caches.jl

@@ -420,7 +420,7 @@ function alg_cache(
     btmp = zero(u)
     bs3cache = BS3Cache(
         u, uprev, bk1, bk2, bk3, bk4, butilde, btmp, batmp, tab,
-        trivial_limiter!, trivial_limiter!, False()
+        trivial_limiter!, trivial_limiter!, Serial()
     )
     fsalfirst = zero(rate_prototype)
     k4 = zero(rate_prototype)
@@ -527,7 +527,7 @@ function alg_cache(
     recursivefill!(ratmp, false)
     rk4cache = RK4Cache(
         u, uprev, rk1, rk2, rk3, rk4, rk, rtmp, ratmp, trivial_limiter!,
-        trivial_limiter!, False()
+        trivial_limiter!, Serial()
     )
     fsalfirst = zero(rate_prototype)
     k4 = zero(rate_prototype)
@@ -636,7 +636,7 @@ function alg_cache(
     recursivefill!(ratmp, false)
     rk4cache = RK4Cache(
         u, uprev, rk1, rk2, rk3, rk4, rk, rtmp, ratmp, trivial_limiter!,
-        trivial_limiter!, False()
+        trivial_limiter!, Serial()
     )
     fsalfirst = zero(rate_prototype)
     k4 = zero(rate_prototype)
@@ -750,7 +750,7 @@ function alg_cache(
     btmp = zero(u)
     bs3cache = BS3Cache(
         u, uprev, bk1, bk2, bk3, bk4, butilde, btmp, batmp, tab,
-        trivial_limiter!, trivial_limiter!, False()
+        trivial_limiter!, trivial_limiter!, Serial()
     )
     fsalfirst = zero(rate_prototype)
     k4 = zero(rate_prototype)
@@ -869,7 +869,7 @@ function alg_cache(
     recursivefill!(ratmp, false)
     rk4cache = RK4Cache(
         u, uprev, rk1, rk2, rk3, rk4, rk, rtmp, ratmp, trivial_limiter!,
-        trivial_limiter!, False()
+        trivial_limiter!, Serial()
     )
     fsalfirst = zero(rate_prototype)
     k4 = zero(rate_prototype)
@@ -988,7 +988,7 @@ function alg_cache(
     recursivefill!(ratmp, false)
     rk4cache = RK4Cache(
         u, uprev, rk1, rk2, rk3, rk4, rk, rtmp, ratmp, trivial_limiter!,
-        trivial_limiter!, False()
+        trivial_limiter!, Serial()
     )
     fsalfirst = zero(rate_prototype)
     k4 = zero(rate_prototype)

lib/OrdinaryDiffEqAdamsBashforthMoulton/src/algorithms.jl

@@ -4,11 +4,11 @@ Problems. Computational Mathematics (2nd revised ed.), Springer (1996) doi:
 https://doi.org/10.1007/978-3-540-78862-1"""
 
 keyword_default_description = """
-- `thread`: determines whether internal broadcasting on appropriate CPU arrays should be serial (`thread = OrdinaryDiffEq.False()`) or use multiple threads (`thread = OrdinaryDiffEq.True()`) when Julia is started with multiple threads.
+- `thread`: determines whether internal broadcasting on appropriate CPU arrays should be serial (`thread = Serial()`) or use multiple threads (`thread = Threaded()`) when Julia is started with multiple threads.
 """
 
 keyword_default = """
-thread = OrdinaryDiffEq.False(),
+thread = Serial(),
 """
 
 @doc generic_solver_docstring(
@@ -21,7 +21,7 @@ thread = OrdinaryDiffEq.False(),
     keyword_default
 )
 Base.@kwdef struct AB3{Thread} <: OrdinaryDiffEqAlgorithm
-    thread::Thread = False()
+    thread::Thread = Serial()
 end
 
 @doc generic_solver_docstring(
@@ -34,7 +34,7 @@ end
     keyword_default
 )
 Base.@kwdef struct AB4{Thread} <: OrdinaryDiffEqAlgorithm
-    thread::Thread = False()
+    thread::Thread = Serial()
 end
 
 @doc generic_solver_docstring(
@@ -47,7 +47,7 @@ end
     keyword_default
 )
 Base.@kwdef struct AB5{Thread} <: OrdinaryDiffEqAlgorithm
-    thread::Thread = False()
+    thread::Thread = Serial()
 end
 
 @doc generic_solver_docstring(
@@ -61,7 +61,7 @@ end
     keyword_default
 )
 Base.@kwdef struct ABM32{Thread} <: OrdinaryDiffEqAlgorithm
-    thread::Thread = False()
+    thread::Thread = Serial()
 end
 
 @doc generic_solver_docstring(
@@ -75,7 +75,7 @@ end
     keyword_default
 )
 Base.@kwdef struct ABM43{Thread} <: OrdinaryDiffEqAlgorithm
-    thread::Thread = False()
+    thread::Thread = Serial()
 end
 
 @doc generic_solver_docstring(
@@ -89,7 +89,7 @@ end
     keyword_default
 )
 Base.@kwdef struct ABM54{Thread} <: OrdinaryDiffEqAlgorithm
-    thread::Thread = False()
+    thread::Thread = Serial()
 end
 
 # Variable Step Size Adams methods
@@ -104,7 +104,7 @@ end
     keyword_default
 )
 Base.@kwdef struct VCAB3{Thread} <: OrdinaryDiffEqAdaptiveAlgorithm
-    thread::Thread = False()
+    thread::Thread = Serial()
 end
 
 @doc generic_solver_docstring(
@@ -117,7 +117,7 @@ end
     keyword_default
 )
 Base.@kwdef struct VCAB4{Thread} <: OrdinaryDiffEqAdaptiveAlgorithm
-    thread::Thread = False()
+    thread::Thread = Serial()
 end
 
 @doc generic_solver_docstring(
@@ -130,7 +130,7 @@ end
     keyword_default
 )
 Base.@kwdef struct VCAB5{Thread} <: OrdinaryDiffEqAdaptiveAlgorithm
-    thread::Thread = False()
+    thread::Thread = Serial()
 end
 
 @doc generic_solver_docstring(
@@ -143,7 +143,7 @@ end
     keyword_default
 )
 Base.@kwdef struct VCABM3{Thread} <: OrdinaryDiffEqAdaptiveAlgorithm
-    thread::Thread = False()
+    thread::Thread = Serial()
 end
 
 @doc generic_solver_docstring(
@@ -156,7 +156,7 @@ end
     keyword_default
 )
 Base.@kwdef struct VCABM4{Thread} <: OrdinaryDiffEqAdaptiveAlgorithm
-    thread::Thread = False()
+    thread::Thread = Serial()
 end
 
 @doc generic_solver_docstring(
@@ -169,7 +169,7 @@ end
     keyword_default
 )
 Base.@kwdef struct VCABM5{Thread} <: OrdinaryDiffEqAdaptiveAlgorithm
-    thread::Thread = False()
+    thread::Thread = Serial()
 end
 
 # Variable Order and Variable Step Size Adams methods
@@ -184,5 +184,5 @@ end
     keyword_default
 )
 Base.@kwdef struct VCABM{Thread} <: OrdinaryDiffEqAdamsVarOrderVarStepAlgorithm
-    thread::Thread = False()
+    thread::Thread = Serial()
 end

lib/OrdinaryDiffEqAdamsBashforthMoulton/test/regression_test_threading.jl

@@ -1,7 +1,7 @@
 using OrdinaryDiffEqAdamsBashforthMoulton, ODEProblemLibrary
 import OrdinaryDiffEqCore: OrdinaryDiffEqAdaptiveAlgorithm
 using Test
-using Static
+using FastBroadcast: Serial, Threaded
 
 algorithms = [
     AB3, AB4, AB5, ABM32, ABM43, ABM54, VCAB3, VCAB4, VCAB5, VCABM3, VCABM4, VCABM5, VCABM,
@@ -12,11 +12,11 @@ problem = ODEProblemLibrary.prob_ode_linear
 @testset "Regression test for threading versions vs non threading versions" begin
     @testset "$ALG" for ALG in algorithms
         if ALG isa OrdinaryDiffEqAdaptiveAlgorithm
-            sol_thread = solve(problem, ALG(Static.True()))
-            sol_nothread = solve(problem, ALG(Static.False()))
+            sol_thread = solve(problem, ALG(Threaded()))
+            sol_nothread = solve(problem, ALG(Serial()))
         else
-            sol_thread = solve(problem, ALG(Static.True()), dt = 1 // 2^9)
-            sol_nothread = solve(problem, ALG(Static.False()), dt = 1 // 2^9)
+            sol_thread = solve(problem, ALG(Threaded()), dt = 1 // 2^9)
+            sol_nothread = solve(problem, ALG(Serial()), dt = 1 // 2^9)
         end
         @test all(sol_nothread .== sol_thread)
     end