ci: taming down CI timings (#903)

* docs: fix links

* test: run tests in parallel

* fix: patch for jacobian

* ci: remove Logging group

* test: remove redundant files

* test: rename file

* ci: remove Logging group

* ci: allow depwarn

* test: fix typo

* ci: cancel intermediate runs

* test: minor fixes

* test: more testing

* test: more test fixes

* test: adjust strategy to reduce runtime

* docs: simplify the examples

.github/workflows/Downgrade.yml

@@ -10,6 +10,11 @@ on:
       - master
     paths-ignore:
       - 'docs/**'
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: ${{ github.ref_name != github.event.repository.default_branch || github.ref != 'refs/tags/v*' }}
+
 jobs:
   test:
     runs-on: ubuntu-latest
@@ -23,7 +28,6 @@ jobs:
           - NNPDE1
           - NNPDE2
           - AdaptiveLoss
-          - Logging
           - Forward
           - DGM
           - NNODE

.github/workflows/Tests.yml

@@ -34,7 +34,6 @@ jobs:
           - "NNPDE1"
           - "NNPDE2"
           - "AdaptiveLoss"
-          - "Logging"
           - "Forward"
           - "DGM"
           - "NNODE"
@@ -45,4 +44,5 @@ jobs:
       group: "${{ matrix.group }}"
       julia-version: "${{ matrix.version }}"
       coverage-directories: "src,ext"
+      julia-runtest-depwarn: "yes"  # TensorBoardLogger has a global depwarn
     secrets: "inherit"

Project.toml

@@ -54,7 +54,7 @@ NeuralPDETensorBoardLoggerExt = "TensorBoardLogger"
 ADTypes = "1.9.0"
 Adapt = "4"
 AdvancedHMC = "0.6.1"
-Aqua = "0.8"
+Aqua = "0.8.9"
 ArrayInterface = "7.11"
 CUDA = "5.5.2"
 ChainRulesCore = "1.24"
@@ -69,7 +69,9 @@ ExplicitImports = "1.10.1"
 Flux = "0.14.22"
 ForwardDiff = "0.10.36"
 Functors = "0.4.12"
+Hwloc = "3.3.0"
 Integrals = "4.5"
+InteractiveUtils = "<0.0.1, 1"
 IntervalSets = "0.7.10"
 LineSearches = "7.3"
 LinearAlgebra = "1.10"
@@ -88,14 +90,13 @@ Optimization = "4"
 OptimizationOptimJL = "0.4"
 OptimizationOptimisers = "0.3"
 OrdinaryDiffEq = "6.87"
-Pkg = "1.10"
 Printf = "1.10"
 QuasiMonteCarlo = "0.3.2"
 Random = "1"
+ReTestItems = "1.29.0"
 RecursiveArrayTools = "3.27.0"
 Reexport = "1.2"
 RuntimeGeneratedFunctions = "0.5.12"
-SafeTestsets = "0.1"
 SciMLBase = "2.56"
 Statistics = "1.10"
 StochasticDiffEq = "6.69.1"
@@ -114,18 +115,19 @@ CUDA = "052768ef-5323-5732-b1bb-66c8b64840ba"
 DiffEqNoiseProcess = "77a26b50-5914-5dd7-bc55-306e6241c503"
 ExplicitImports = "7d51a73a-1435-4ff3-83d9-f097790105c7"
 Flux = "587475ba-b771-5e3f-ad9e-33799f191a9c"
+Hwloc = "0e44f5e4-bd66-52a0-8798-143a42290a1d"
+InteractiveUtils = "b77e0a4c-d291-57a0-90e8-8db25a27a240"
 LineSearches = "d3d80556-e9d4-5f37-9878-2ab0fcc64255"
 LuxCUDA = "d0bbae9a-e099-4d5b-a835-1c6931763bda"
 LuxCore = "bb33d45b-7691-41d6-9220-0943567d0623"
 LuxLib = "82251201-b29d-42c6-8e01-566dec8acb11"
 MethodOfLines = "94925ecb-adb7-4558-8ed8-f975c56a0bf4"
 OptimizationOptimJL = "36348300-93cb-4f02-beb5-3c3902f8871e"
 OrdinaryDiffEq = "1dea7af3-3e70-54e6-95c3-0bf5283fa5ed"
-Pkg = "44cfe95a-1eb2-52ea-b672-e2afdf69b78f"
-SafeTestsets = "1bc83da4-3b8d-516f-aca4-4fe02f6d838f"
+ReTestItems = "817f1d60-ba6b-4fd5-9520-3cf149f6a823"
 StochasticDiffEq = "789caeaf-c7a9-5a7d-9973-96adeb23e2a0"
 TensorBoardLogger = "899adc3e-224a-11e9-021f-63837185c80f"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 
 [targets]
-test = ["Aqua", "CUDA", "DiffEqNoiseProcess", "ExplicitImports", "Flux", "LineSearches", "LuxCUDA", "LuxCore", "LuxLib", "MethodOfLines", "OptimizationOptimJL", "OrdinaryDiffEq", "Pkg", "SafeTestsets", "StochasticDiffEq", "TensorBoardLogger", "Test"]
+test = ["Aqua", "CUDA", "DiffEqNoiseProcess", "ExplicitImports", "Flux", "Hwloc", "InteractiveUtils", "LineSearches", "LuxCUDA", "LuxCore", "LuxLib", "MethodOfLines", "OptimizationOptimJL", "OrdinaryDiffEq", "ReTestItems", "StochasticDiffEq", "TensorBoardLogger", "Test"]

docs/make.jl

@@ -18,5 +18,4 @@ makedocs(sitename = "NeuralPDE.jl",
         canonical = "https://docs.sciml.ai/NeuralPDE/stable/"),
     pages = pages)
 
-deploydocs(repo = "github.com/SciML/NeuralPDE.jl.git";
-    push_preview = true)
+deploydocs(repo = "github.com/SciML/NeuralPDE.jl.git"; push_preview = true)

docs/src/index.md

@@ -17,7 +17,7 @@ networks which both approximate physical laws and real data simultaneously.
   - Specialized forms for solving `ODEProblem`s with neural networks.
   - Compatibility with [Flux.jl](https://fluxml.ai/) and [Lux.jl](https://lux.csail.mit.edu/).
     for all the GPU-powered machine learning layers available from those libraries.
-  - Compatibility with [NeuralOperators.jl](https://docs.sciml.ai/NeuralOperators/stable/) for
+  - Compatibility with [NeuralOperators.jl](https://github.com/SciML/NeuralOperators.jl) for
     mixing DeepONets and other neural operators (Fourier Neural Operators, Graph Neural Operators,
     etc.) with physics-informed loss functions.
 

docs/src/tutorials/derivative_neural_network.md

@@ -52,9 +52,8 @@ We approximate the derivative of the neural network with another neural network
 using the second numeric derivative `Dt(Dtu1(t,x))`.
 
 ```@example derivativenn
-using NeuralPDE, Lux, ModelingToolkit
-using Optimization, OptimizationOptimisers, OptimizationOptimJL, LineSearches
-using Plots
+using NeuralPDE, Lux, ModelingToolkit, Optimization, OptimizationOptimisers,
+      OptimizationOptimJL, LineSearches, Plots
 using ModelingToolkit: Interval, infimum, supremum
 
 @parameters t, x
@@ -63,35 +62,40 @@ Dx = Differential(x)
 @variables u1(..), u2(..), u3(..)
 @variables Dxu1(..) Dtu1(..) Dxu2(..) Dtu2(..)
 
-eqs_ = [Dt(Dtu1(t, x)) ~ Dx(Dxu1(t, x)) + u3(t, x) * sin(pi * x),
-    Dt(Dtu2(t, x)) ~ Dx(Dxu2(t, x)) + u3(t, x) * cos(pi * x),
-    exp(-t) ~ u1(t, x) * sin(pi * x) + u2(t, x) * cos(pi * x)]
-
-bcs_ = [u1(0.0, x) ~ sin(pi * x),
-    u2(0.0, x) ~ cos(pi * x),
-    Dt(u1(0, x)) ~ -sin(pi * x),
-    Dt(u2(0, x)) ~ -cos(pi * x),
+eqs_ = [
+    Dt(Dtu1(t, x)) ~ Dx(Dxu1(t, x)) + u3(t, x) * sinpi(x),
+    Dt(Dtu2(t, x)) ~ Dx(Dxu2(t, x)) + u3(t, x) * cospi(x),
+    exp(-t) ~ u1(t, x) * sinpi(x) + u2(t, x) * cospi(x)
+]
+
+bcs_ = [
+    u1(0.0, x) ~ sinpi(x),
+    u2(0.0, x) ~ cospi(x),
+    Dt(u1(0, x)) ~ -sinpi(x),
+    Dt(u2(0, x)) ~ -cospi(x),
     u1(t, 0.0) ~ 0.0,
     u2(t, 0.0) ~ exp(-t),
     u1(t, 1.0) ~ 0.0,
-    u2(t, 1.0) ~ -exp(-t)]
+    u2(t, 1.0) ~ -exp(-t)
+]
 
-der_ = [Dt(u1(t, x)) ~ Dtu1(t, x),
+der_ = [
+    Dt(u1(t, x)) ~ Dtu1(t, x),
     Dt(u2(t, x)) ~ Dtu2(t, x),
     Dx(u1(t, x)) ~ Dxu1(t, x),
-    Dx(u2(t, x)) ~ Dxu2(t, x)]
+    Dx(u2(t, x)) ~ Dxu2(t, x)
+]
 
 bcs__ = [bcs_; der_]
 
 # Space and time domains
-domains = [t ∈ Interval(0.0, 1.0),
-    x ∈ Interval(0.0, 1.0)]
+domains = [t ∈ Interval(0.0, 1.0), x ∈ Interval(0.0, 1.0)]
 
 input_ = length(domains)
 n = 15
-chain = [Lux.Chain(Dense(input_, n, Lux.σ), Dense(n, n, Lux.σ), Dense(n, 1)) for _ in 1:7]
+chain = [Chain(Dense(input_, n, σ), Dense(n, n, σ), Dense(n, 1)) for _ in 1:7]
 
-training_strategy = QuadratureTraining(; batch = 200, reltol = 1e-6, abstol = 1e-6)
+training_strategy = StochasticTraining(128)
 discretization = PhysicsInformedNN(chain, training_strategy)
 
 vars = [u1(t, x), u2(t, x), u3(t, x), Dxu1(t, x), Dtu1(t, x), Dxu2(t, x), Dtu2(t, x)]
@@ -126,13 +130,13 @@ using Plots
 ts, xs = [infimum(d.domain):0.01:supremum(d.domain) for d in domains]
 minimizers_ = [res.u.depvar[sym_prob.depvars[i]] for i in 1:length(chain)]
 
-u1_real(t, x) = exp(-t) * sin(pi * x)
-u2_real(t, x) = exp(-t) * cos(pi * x)
+u1_real(t, x) = exp(-t) * sinpi(x)
+u2_real(t, x) = exp(-t) * cospi(x)
 u3_real(t, x) = (1 + pi^2) * exp(-t)
-Dxu1_real(t, x) = pi * exp(-t) * cos(pi * x)
-Dtu1_real(t, x) = -exp(-t) * sin(pi * x)
-Dxu2_real(t, x) = -pi * exp(-t) * sin(pi * x)
-Dtu2_real(t, x) = -exp(-t) * cos(pi * x)
+Dxu1_real(t, x) = pi * exp(-t) * cospi(x)
+Dtu1_real(t, x) = -exp(-t) * sinpi(x)
+Dxu2_real(t, x) = -pi * exp(-t) * sinpi(x)
+Dtu2_real(t, x) = -exp(-t) * cospi(x)
 
 function analytic_sol_func_all(t, x)
     [u1_real(t, x), u2_real(t, x), u3_real(t, x),