Update benchmarks to OrdinaryDiffEq v7 / SciMLBase v3 / RecursiveArrayTools v4 stack

benchmarks/AdaptiveSDE/Project.toml

@@ -6,6 +6,7 @@ Plots = "91a5bcdd-55d7-5caf-9e0b-520d859cae80"
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
 SDEProblemLibrary = "c72e72a9-a271-4b2b-8966-303ed956772e"
 SciMLBenchmarks = "31c91b34-3c75-11e9-0341-95557aab0344"
+SciMLLogging = "a6db7da4-7206-11f0-1eab-35f2a5dbe1d1"
 StochasticDiffEq = "789caeaf-c7a9-5a7d-9973-96adeb23e2a0"
 
 [compat]
@@ -14,4 +15,5 @@ ParallelDataTransfer = "0.5"
 Plots = "1.4"
 SDEProblemLibrary = "0.1, 1"
 SciMLBenchmarks = "0.1"
-StochasticDiffEq = "6.23"
+SciMLLogging = "1, 2"
+StochasticDiffEq = "7"

benchmarks/AdaptiveSDE/qmaxDetermination.jmd

@@ -10,6 +10,7 @@ means = Array{Float64}(undef, length(qs), 4)
 
 using StochasticDiffEq, SDEProblemLibrary, Random,
       Plots, ParallelDataTransfer, DiffEqMonteCarlo, Distributed
+using SciMLLogging
 Random.seed!(99)
 
 full_prob = SDEProblemLibrary.oval2ModelExample(largeFluctuations = true, useBigs = false)
@@ -41,7 +42,7 @@ println("Setup Complete")
 
 function runAdaptive(i, k)
     sol = solve(prob, SRIW1(), dt = 1/2^(8), abstol = 2.0^(-15), reltol = 2.0^(-10),
-        verbose = false, maxIters = Int(1e12), qmax = qs[k])
+        verbose = SciMLLogging.None(), maxIters = Int(1e12), qmax = qs[k])
     Int(any(isnan, sol[end]) || sol.t[end] != 1)
 end
 

benchmarks/AstroChem/Project.toml

@@ -1,18 +1,24 @@
 [compat]
 BenchmarkTools = "1"
-Catalyst = "15"
-DiffEqDevTools = "2"
+Catalyst = "16"
+DiffEqDevTools = "3"
 LSODA = "0.7"
 LinearSolve = "3"
 NaNMath = "1"
 ODEInterface = "0.5"
-ODEInterfaceDiffEq = "3"
-OrdinaryDiffEq = "6"
-OrdinaryDiffEqCore = "3"
+ODEInterfaceDiffEq = "4"
+OrdinaryDiffEq = "7"
+OrdinaryDiffEqBDF = "2"
+OrdinaryDiffEqCore = "4"
+OrdinaryDiffEqExtrapolation = "2"
+OrdinaryDiffEqFIRK = "2"
+OrdinaryDiffEqRosenbrock = "2"
+OrdinaryDiffEqSDIRK = "2"
 Plots = "1"
 RecursiveFactorization = "0.2"
-Sundials = "5"
-Symbolics = "6"
+SciMLLogging = "1, 2"
+Sundials = "6"
+Symbolics = "7"
 
 [deps]
 BenchmarkTools = "6e4b80f9-dd63-53aa-95a3-0cdb28fa8baf"
@@ -24,8 +30,14 @@ NaNMath = "77ba4419-2d1f-58cd-9bb1-8ffee604a2e3"
 ODEInterface = "54ca160b-1b9f-5127-a996-1867f4bc2a2c"
 ODEInterfaceDiffEq = "09606e27-ecf5-54fc-bb29-004bd9f985bf"
 OrdinaryDiffEq = "1dea7af3-3e70-54e6-95c3-0bf5283fa5ed"
+OrdinaryDiffEqBDF = "6ad6398a-0878-4a85-9266-38940aa047c8"
 OrdinaryDiffEqCore = "bbf590c4-e513-4bbe-9b18-05decba2e5d8"
+OrdinaryDiffEqExtrapolation = "becaefa8-8ca2-5cf9-886d-c06f3d2bd2c4"
+OrdinaryDiffEqFIRK = "5960d6e9-dd7a-4743-88e7-cf307b64f125"
+OrdinaryDiffEqRosenbrock = "43230ef6-c299-4910-a778-202eb28ce4ce"
+OrdinaryDiffEqSDIRK = "2d112036-d095-4a1e-ab9a-08536f3ecdbf"
 Plots = "91a5bcdd-55d7-5caf-9e0b-520d859cae80"
 RecursiveFactorization = "f2c3362d-daeb-58d1-803e-2bc74f2840b4"
+SciMLLogging = "a6db7da4-7206-11f0-1eab-35f2a5dbe1d1"
 Sundials = "c3572dad-4567-51f8-b174-8c6c989267f4"
 Symbolics = "0c5d862f-8b57-4792-8d23-62f2024744c7"

benchmarks/AstroChem/astrochem.jmd

@@ -5,7 +5,9 @@ author: Gijs Vermariën and Chris Rackauckas
 
 ```julia
 using Catalyst
+using SciMLLogging
 using OrdinaryDiffEq
+using OrdinaryDiffEqBDF, OrdinaryDiffEqExtrapolation, OrdinaryDiffEqFIRK, OrdinaryDiffEqRosenbrock, OrdinaryDiffEqSDIRK
 using Plots
 using Symbolics
 using DiffEqDevTools
@@ -329,7 +331,7 @@ setups = [
     #Dict(:alg=>ImplicitEulerBarycentricExtrapolation(min_order = 5, threading = OrdinaryDiffEqCore.PolyesterThreads())),
     Dict(:alg=>ImplicitEulerBarycentricExtrapolation(min_order = 5, threading = false))
 ]
-wp = WorkPrecisionSet(oprob, abstols, reltols, setups; verbose = false,
+wp = WorkPrecisionSet(oprob, abstols, reltols, setups; verbose = SciMLLogging.None(),
     save_everystep = false, appxsol = refsol, maxiters = Int(1e5), numruns = 10)
 plot(wp)
 ```
@@ -420,7 +422,7 @@ setups = [
     #Dict(:alg=>ImplicitEulerBarycentricExtrapolation(min_order = 5, threading = OrdinaryDiffEqCore.PolyesterThreads())),
     #Dict(:alg=>ImplicitEulerBarycentricExtrapolation(min_order = 5, threading = false)),
 ]
-wp = WorkPrecisionSet(oprob, abstols, reltols, setups; verbose = false,
+wp = WorkPrecisionSet(oprob, abstols, reltols, setups; verbose = SciMLLogging.None(),
     save_everystep = false, appxsol = refsol, maxiters = Int(1e5), numruns = 10,
     print_names = true)
 plot(wp)

benchmarks/AstroChem/nelson.jmd

@@ -5,6 +5,7 @@ author: Nina De La Torre (Advised by Stella Offner) and Chris Rackauckas
 
 ```julia
 using OrdinaryDiffEq
+using OrdinaryDiffEqBDF, OrdinaryDiffEqExtrapolation, OrdinaryDiffEqFIRK, OrdinaryDiffEqRosenbrock, OrdinaryDiffEqSDIRK
 using DiffEqDevTools, Plots
 using Sundials, LSODA
 using ODEInterface, ODEInterfaceDiffEq

benchmarks/AutomaticDifferentiation/BrussScaling.jmd

@@ -7,6 +7,7 @@ From the paper [A Comparison of Automatic Differentiation and Continuous Sensiti
 
 ```julia
 using OrdinaryDiffEq, ReverseDiff, ForwardDiff, FiniteDiff, SciMLSensitivity
+using OrdinaryDiffEqRosenbrock
 using LinearAlgebra, Tracker, Mooncake, Plots
 ```
 
@@ -216,17 +217,17 @@ makebrusselator!(dict, n) = get!(()->makebrusselator(n), dict, n)
 _adjoint_methods_iq = ntuple(2) do ii
     Alg = (InterpolatingAdjoint, QuadratureAdjoint)[ii]
     (
-        user = Alg(autodiff = false, autojacvec = false), # user Jacobian
-        adjc = Alg(autodiff = true, autojacvec = false), # AD Jacobian
-        advj = Alg(autodiff = true, autojacvec = EnzymeVJP()) # AD vJ
+        user = Alg(autodiff = AutoFiniteDiff(), autojacvec = false), # user Jacobian
+        adjc = Alg(autodiff = AutoForwardDiff(), autojacvec = false), # AD Jacobian
+        advj = Alg(autodiff = AutoForwardDiff(), autojacvec = EnzymeVJP()) # AD vJ
     )
 end |> NamedTuple{(:interp, :quad)}
-# GaussAdjoint/GaussKronrodAdjoint do not support user-provided Jacobians (autodiff=false)
+# GaussAdjoint/GaussKronrodAdjoint do not support user-provided Jacobians (autodiff=AutoFiniteDiff())
 _adjoint_methods_g = ntuple(2) do ii
     Alg = (GaussAdjoint, GaussKronrodAdjoint)[ii]
     (
-        adjc = Alg(autodiff = true, autojacvec = false), # AD Jacobian
-        advj = Alg(autodiff = true, autojacvec = EnzymeVJP()) # AD vJ
+        adjc = Alg(autodiff = AutoForwardDiff(), autojacvec = false), # AD Jacobian
+        advj = Alg(autodiff = AutoForwardDiff(), autojacvec = EnzymeVJP()) # AD vJ
     )
 end |> NamedTuple{(:gauss, :gausskronrod)}
 @isdefined(ADJOINT_METHODS_IQ) ||
@@ -275,8 +276,8 @@ end
 #=
 reversediff = map(reversediffn) do n
   bfun, b_u0, b_p, brusselator_jac, brusselator_comp = makebrusselator!(PROBS, n)
-  @elapsed auto_sen_l2(bfun, b_u0, tspan, b_p, bt, (Rodas5(autodiff=false)); diffalg=(ReverseDiff.gradient), tols...)
-  t = @elapsed auto_sen_l2(bfun, b_u0, tspan, b_p, bt, (Rodas5(autodiff=false)); diffalg=(ReverseDiff.gradient), tols...)
+  @elapsed auto_sen_l2(bfun, b_u0, tspan, b_p, bt, (Rodas5(autodiff=AutoFiniteDiff())); diffalg=(ReverseDiff.gradient), tols...)
+  t = @elapsed auto_sen_l2(bfun, b_u0, tspan, b_p, bt, (Rodas5(autodiff=AutoFiniteDiff())); diffalg=(ReverseDiff.gradient), tols...)
   @show n,t
   t
 end
@@ -301,7 +302,7 @@ is complete before we start timing.
 ```julia
 let n = first(csan)
     bfun, b_u0, b_p, brusselator_jac, brusselator_comp = makebrusselator!(PROBS, n)
-    solver = Rodas5(autodiff = false)
+    solver = Rodas5(autodiff = AutoFiniteDiff())
     for alg in ADJOINT_METHODS_IQ
         f = SciMLSensitivity.alg_autodiff(alg) ? bfun :
             ODEFunction(bfun, jac = brusselator_jac)
@@ -320,7 +321,7 @@ csa_iq = map(csan) do n
         @info "Running $alg"
         f = SciMLSensitivity.alg_autodiff(alg) ? bfun :
             ODEFunction(bfun, jac = brusselator_jac)
-        solver = Rodas5(autodiff = false)
+        solver = Rodas5(autodiff = AutoFiniteDiff())
         @time diffeq_sen_l2(f, b_u0, tspan, b_p, bt, solver; sensalg = alg, tols...)
         t = @elapsed diffeq_sen_l2(f, b_u0, tspan, b_p, bt, solver; sensalg = alg, tols...)
         return t
@@ -335,7 +336,7 @@ csa_g = map(csan) do n
     bfun, b_u0, b_p, brusselator_jac, brusselator_comp = makebrusselator!(PROBS, n)
     @time ts = map(ADJOINT_METHODS_G) do alg
         @info "Running $alg"
-        solver = Rodas5(autodiff = false)
+        solver = Rodas5(autodiff = AutoFiniteDiff())
         @time diffeq_sen_l2(bfun, b_u0, tspan, b_p, bt, solver; sensalg = alg, tols...)
         t = @elapsed diffeq_sen_l2(bfun, b_u0, tspan, b_p, bt, solver; sensalg = alg, tols...)
         return t
@@ -398,10 +399,10 @@ tols = (abstol = 1e-5, reltol = 1e-7)
 _adjoint_methods = ntuple(4) do ii
     Alg = (InterpolatingAdjoint, QuadratureAdjoint, GaussAdjoint, GaussKronrodAdjoint)[ii]
     (
-        advj1 = Alg(autodiff = true, autojacvec = EnzymeVJP()), # AD vJ (Enzyme)
-        advj2 = Alg(autodiff = true, autojacvec = ReverseDiffVJP(false)), # AD vJ (ReverseDiff)
-        advj3 = Alg(autodiff = true, autojacvec = ReverseDiffVJP(true)), # AD vJ (Compiled ReverseDiff)
-        advj4 = Alg(autodiff = true, autojacvec = SciMLSensitivity.MooncakeVJP()) # AD vJ (Mooncake)
+        advj1 = Alg(autodiff = AutoForwardDiff(), autojacvec = EnzymeVJP()), # AD vJ (Enzyme)
+        advj2 = Alg(autodiff = AutoForwardDiff(), autojacvec = ReverseDiffVJP(false)), # AD vJ (ReverseDiff)
+        advj3 = Alg(autodiff = AutoForwardDiff(), autojacvec = ReverseDiffVJP(true)), # AD vJ (Compiled ReverseDiff)
+        advj4 = Alg(autodiff = AutoForwardDiff(), autojacvec = SciMLSensitivity.MooncakeVJP()) # AD vJ (Mooncake)
     )
 end |> NamedTuple{(:interp, :quad, :gauss, :gausskronrod)}
 adjoint_methods = mapreduce(collect, vcat, _adjoint_methods)
@@ -412,7 +413,7 @@ Warmup: compile all VJP backends before benchmarking.
 ```julia
 let n = first(csan)
     bfun, b_u0, b_p, brusselator_jac, brusselator_comp = makebrusselator!(PROBS, n)
-    solver = Rodas5(autodiff = false)
+    solver = Rodas5(autodiff = AutoFiniteDiff())
     for alg in adjoint_methods
         f = SciMLSensitivity.alg_autodiff(alg) ? bfun :
             ODEFunction(bfun, jac = brusselator_jac)
@@ -428,7 +429,7 @@ csavjp = map(csan) do n
         @info "Running $alg"
         f = SciMLSensitivity.alg_autodiff(alg) ? bfun :
             ODEFunction(bfun, jac = brusselator_jac)
-        solver = Rodas5(autodiff = false)
+        solver = Rodas5(autodiff = AutoFiniteDiff())
         @time diffeq_sen_l2(f, b_u0, tspan, b_p, bt, solver; sensalg = alg, tols...)
         t = @elapsed diffeq_sen_l2(f, b_u0, tspan, b_p, bt, solver; sensalg = alg, tols...)
         return t
@@ -738,17 +739,17 @@ end
 ```julia
 adjoint_ad_configs = [
     ("Interp user-Jacobian",
-     "InterpolatingAdjoint(autodiff = false, autojacvec = false)", true),
+     "InterpolatingAdjoint(autodiff = AutoFiniteDiff(), autojacvec = false)", true),
     ("Interp AD-Jacobian",
-     "InterpolatingAdjoint(autodiff = true, autojacvec = false)", false),
+     "InterpolatingAdjoint(autodiff = AutoForwardDiff(), autojacvec = false)", false),
     ("Quad user-Jacobian",
-     "QuadratureAdjoint(autodiff = false, autojacvec = false)", true),
+     "QuadratureAdjoint(autodiff = AutoFiniteDiff(), autojacvec = false)", true),
     ("Quad AD-Jacobian",
-     "QuadratureAdjoint(autodiff = true, autojacvec = false)", false),
+     "QuadratureAdjoint(autodiff = AutoForwardDiff(), autojacvec = false)", false),
     ("Gauss AD-Jacobian",
-     "GaussAdjoint(autodiff = true, autojacvec = false)", false),
+     "GaussAdjoint(autodiff = AutoForwardDiff(), autojacvec = false)", false),
     ("GaussKronrod AD-Jacobian",
-     "GaussKronrodAdjoint(autodiff = true, autojacvec = false)", false),
+     "GaussKronrodAdjoint(autodiff = AutoForwardDiff(), autojacvec = false)", false),
 ]
 
 adjoint_ad_mem = map(adjoint_ad_configs) do (name, sensalg_str, needs_jac)
@@ -757,7 +758,7 @@ adjoint_ad_mem = map(adjoint_ad_configs) do (name, sensalg_str, needs_jac)
         result = run_memory_benchmark(n, """
         sensalg = $(sensalg_str)
         f = $(f_expr)
-        solver = Rodas5(autodiff = false)
+        solver = Rodas5(autodiff = AutoFiniteDiff())
         diffeq_sen_l2(f, b_u0, tspan, b_p, bt, solver; sensalg = sensalg, tols...)
         t = @elapsed diffeq_sen_l2(f, b_u0, tspan, b_p, bt, solver;
             sensalg = sensalg, tols...)