Daniel Karrasch

@nanosoldier `runtests()`

Let's see what the pkgeval run says, but I think we should/could slowly start to collectively think about whether we want to do this, or what should be taken care...

@nanosoldier `runtests(["FloatTracker", "EulerAngles", "NumericalAlgorithms", "ConvexHulls2d", "SubSIt", "CompressedSparseBlocks", "MLKernels", "NMF", "VibrationGEPHelpers", "JSOSolvers", "ExtendableSparse", "EcologicalNetworksPlots", "CrystalNets", "OptimalPortfolios", "GLPK", "PlantRayTracer", "ProximalOperators", "Gtk4", "SkyDomes", "BasisMatrices", "TaylorIntegration", "TransitionsInTimeseries", "MixedModels", "ClimaCoreSpectra", "ClimaCorePlots", "LowLevelParticleFilters", "MultiStateSystems", "ManifoldDiffEq", "ConScape",...

@nanosoldier `runtests(["FloatTracker", "EulerAngles", "NumericalAlgorithms", "ConvexHulls2d", "SubSIt", "CompressedSparseBlocks", "MLKernels", "NMF", "VibrationGEPHelpers", "JSOSolvers", "ExtendableSparse", "EcologicalNetworksPlots", "CrystalNets", "OptimalPortfolios", "GLPK", "PlantRayTracer", "ProximalOperators", "Gtk4", "SkyDomes", "BasisMatrices", "TaylorIntegration", "TransitionsInTimeseries", "MixedModels", "ClimaCoreSpectra", "ClimaCorePlots", "LowLevelParticleFilters", "MultiStateSystems", "ManifoldDiffEq", "ConScape",...

@nanosoldier `runtests()`

@nanosoldier `runtests(["SubSIt", "ReduceWindows", "CompressedSparseBlocks", "EasyCurl", "IterativeSolvers", "LinearMaps", "VibrationGEPHelpers", "PreallocationTools", "ParameterizedQuantumControl", "CrystalNets", "MRICoilSensitivities", "CDDLib", "TaylorIntegration", "BasisMatrices", "ReservoirComputing", "MixedModels", "StructuralEquationModels", "LongwaveModePropagator", "HierarchicalGaussianFiltering", "LowRankIntegrators", "MimiRFFSPs", "Plots", "Knockoffs", "ConceptualClimateModels", "AstrodynamicalModels", "MinimallyDisruptiveCurves", "Biofilm", "SMLMSim"])` Most...

A few packages throw some signal in the most generic multiplication kernel, in the simd-loop applying `muladd`. It could be that this is related to sparse arrays, which don't get...

The following compile time benchmark from @chriselrod ```julia using ForwardDiff, LinearAlgebra d(x, n) = ForwardDiff.Dual(x, ntuple(_ -> randn(), n)) function dualify(A, n, j) if n > 0 A = d.(A,...

@nanosoldier `runbenchmarks("linalg", vs = ":master")`

@nanosoldier `runbenchmarks("linalg", vs = ":release-v1.11")`