Support ForwardDiff 1.0

[ErikQQY]

Let's see if the upgrade breaks anything

[github-actions]

Benchmark Results

	master	51b362f...	master / 51b362f...
Simple Pendulum/IIP/BoundaryValueDiffEqMIRK.MIRK2()	7.71 ± 0.22 ms	7.87 ± 0.26 ms	0.979
Simple Pendulum/IIP/BoundaryValueDiffEqMIRK.MIRK3()	2.87 ± 0.18 ms	2.89 ± 0.19 ms	0.991
Simple Pendulum/IIP/BoundaryValueDiffEqMIRK.MIRK4()	0.997 ± 0.082 ms	1 ± 0.077 ms	0.993
Simple Pendulum/IIP/BoundaryValueDiffEqMIRK.MIRK5()	2.55 ± 0.45 ms	2.58 ± 0.46 ms	0.988
Simple Pendulum/IIP/BoundaryValueDiffEqMIRK.MIRK6()	1.17 ± 0.15 ms	1.17 ± 0.15 ms	0.999
Simple Pendulum/IIP/MultipleShooting(10, Tsit5; grid_coarsening = false)	1.89 ± 0.66 ms	1.9 ± 0.66 ms	0.998
Simple Pendulum/IIP/MultipleShooting(10, Tsit5; grid_coarsening = true)	3.16 ± 1 ms	3.15 ± 1 ms	1
Simple Pendulum/IIP/MultipleShooting(100, Tsit5; grid_coarsening = false)	0.0708 ± 0.02 s	0.0709 ± 0.021 s	0.999
Simple Pendulum/IIP/MultipleShooting(100, Tsit5; grid_coarsening = true)	0.085 ± 0.022 s	0.0841 ± 0.022 s	1.01
Simple Pendulum/IIP/Shooting(Tsit5())	0.265 ± 0.085 ms	0.257 ± 0.082 ms	1.03
Simple Pendulum/OOP/BoundaryValueDiffEqMIRK.MIRK2()	9.1 ± 0.41 ms	9.09 ± 0.36 ms	1
Simple Pendulum/OOP/BoundaryValueDiffEqMIRK.MIRK3()	3.33 ± 0.15 ms	3.34 ± 0.15 ms	0.998
Simple Pendulum/OOP/BoundaryValueDiffEqMIRK.MIRK4()	1.16 ± 0.088 ms	1.16 ± 0.095 ms	1
Simple Pendulum/OOP/BoundaryValueDiffEqMIRK.MIRK5()	2.98 ± 0.48 ms	2.97 ± 0.38 ms	1
Simple Pendulum/OOP/BoundaryValueDiffEqMIRK.MIRK6()	1.39 ± 0.13 ms	1.39 ± 0.14 ms	0.997
Simple Pendulum/OOP/MultipleShooting(10, Tsit5; grid_coarsening = false)	3.45 ± 2.9 ms	3.49 ± 3.1 ms	0.99
Simple Pendulum/OOP/MultipleShooting(10, Tsit5; grid_coarsening = true)	5.85 ± 4.9 ms	5.94 ± 5.5 ms	0.985
Simple Pendulum/OOP/MultipleShooting(100, Tsit5; grid_coarsening = false)	0.122 ± 0.0097 s	0.121 ± 0.017 s	1.02
Simple Pendulum/OOP/MultipleShooting(100, Tsit5; grid_coarsening = true)	0.148 ± 0.024 s	0.149 ± 0.017 s	0.992
Simple Pendulum/OOP/Shooting(Tsit5())	0.637 ± 0.048 ms	0.642 ± 0.065 ms	0.993
time_to_load	4.97 ± 0.018 s	4.97 ± 0.009 s	1

Benchmark Plots

A plot of the benchmark results have been uploaded as an artifact to the workflow run for this PR.
Go to "Actions"->"Benchmark a pull request"->[the most recent run]->"Artifacts" (at the bottom).

[devmotion]

Tests (eg of Core) did not run with ForwardDiff 1. ForwardDiff 1 seems untested.

[ErikQQY]

Indeed, we first need ForwardDiff 1 on upstream packages like NonlinearSolve in SciML/NonlinearSolve.jl#561