feat: add stack support

[avik-pal]

needs testing

[github-actions]

Benchmark Results

	main	dc07f74...	main/dc07f74c34bdb0...
comptime/NN/ViT base (optimize = :after_enzyme)	10.7 s	10.6 s	1.01
comptime/NN/ViT base (optimize = :all)	11 s	10.9 s	1.01
comptime/NN/ViT base (optimize = :before_enzyme)	10.8 s	10.4 s	1.04
comptime/NN/ViT base (optimize = :only_enzyme)	11 s	11 s	0.996
comptime/NN/ViT tiny (optimize = :after_enzyme)	8.37 s	8.17 s	1.02
comptime/NN/ViT tiny (optimize = :all)	8.59 s	8.53 s	1.01
comptime/NN/ViT tiny (optimize = :before_enzyme)	8.48 s	8.56 s	0.991
comptime/NN/ViT tiny (optimize = :only_enzyme)	8.64 s	8.59 s	1.01
comptime/NN/vgg11 bn=false (optimize = :after_enzyme)	1.28 ± 0.036 s	1.33 ± 0.025 s	0.966
comptime/NN/vgg11 bn=false (optimize = :all)	1.14 ± 0.019 s	1.43 ± 0.038 s	0.794
comptime/NN/vgg11 bn=false (optimize = :before_enzyme)	1.1 ± 0.04 s	1.28 ± 0.01 s	0.856
comptime/NN/vgg11 bn=false (optimize = :only_enzyme)	1.14 ± 0.03 s	1.13 ± 0.016 s	1.01
comptime/NN/vgg11 bn=true (optimize = :after_enzyme)	2.14 ± 0.0078 s	2.09 ± 0.084 s	1.03
comptime/NN/vgg11 bn=true (optimize = :all)	2.18 ± 0.024 s	2.16 ± 0.025 s	1.01
comptime/NN/vgg11 bn=true (optimize = :before_enzyme)	2.26 ± 0.032 s	2.18 ± 0.085 s	1.04
comptime/NN/vgg11 bn=true (optimize = :only_enzyme)	2.2 ± 0.08 s	2.11 ± 0.049 s	1.04
comptime/NN/vgg13 bn=false (optimize = :after_enzyme)	1.59 ± 0.12 s	1.62 ± 0.087 s	0.981
comptime/NN/vgg13 bn=false (optimize = :all)	1.75 ± 0.09 s	1.77 ± 0.1 s	0.989
comptime/NN/vgg13 bn=false (optimize = :before_enzyme)	1.74 ± 0.029 s	1.63 ± 0.12 s	1.07
comptime/NN/vgg13 bn=false (optimize = :only_enzyme)	1.75 ± 0.046 s	1.59 ± 0.12 s	1.1
comptime/NN/vgg13 bn=true (optimize = :after_enzyme)	2.92 s	2.81 s	1.04
comptime/NN/vgg13 bn=true (optimize = :all)	2.75 s	2.78 s	0.989
comptime/NN/vgg13 bn=true (optimize = :before_enzyme)	2.98 s	2.81 s	1.06
comptime/NN/vgg13 bn=true (optimize = :only_enzyme)	3.1 s	2.88 s	1.08
comptime/NN/vgg16 bn=false (optimize = :after_enzyme)	2.21 ± 0.0031 s	1.93 ± 0.029 s	1.15
comptime/NN/vgg16 bn=false (optimize = :all)	2.11 ± 0.11 s	2.25 ± 0.011 s	0.938
comptime/NN/vgg16 bn=false (optimize = :before_enzyme)	2.13 ± 0.0098 s	1.91 ± 0.015 s	1.12
comptime/NN/vgg16 bn=false (optimize = :only_enzyme)	2.27 ± 0.0086 s	2.21 ± 0.051 s	1.03
comptime/NN/vgg16 bn=true (optimize = :after_enzyme)	3.59 s	3.39 s	1.06
comptime/NN/vgg16 bn=true (optimize = :all)	3.54 s	3.39 s	1.04
comptime/NN/vgg16 bn=true (optimize = :before_enzyme)	3.57 s	3.44 s	1.04
comptime/NN/vgg16 bn=true (optimize = :only_enzyme)	3.44 s	3.44 s	1
comptime/NN/vgg19 bn=false (optimize = :after_enzyme)	2.31 ± 0.0073 s	2.07 ± 0.017 s	1.12
comptime/NN/vgg19 bn=false (optimize = :all)	2.5 s	2.58 s	0.971
comptime/NN/vgg19 bn=false (optimize = :before_enzyme)	2.58 s	2.25 ± 0.037 s	1.15
comptime/NN/vgg19 bn=false (optimize = :only_enzyme)	2.54 s	2.05 ± 0.016 s	1.24
comptime/NN/vgg19 bn=true (optimize = :after_enzyme)	4.09 s	3.94 s	1.04
comptime/NN/vgg19 bn=true (optimize = :all)	4.31 s	3.96 s	1.09
comptime/NN/vgg19 bn=true (optimize = :before_enzyme)	4.55 s	4.15 s	1.1
comptime/NN/vgg19 bn=true (optimize = :only_enzyme)	4.19 s	4.16 s	1.01
comptime/basics/2D sum (optimize = :after_enzyme)	30.9 ± 2.4 ms	0.0326 ± 0.0028 s	0.948
comptime/basics/2D sum (optimize = :all)	0.0349 ± 0.0029 s	0.0371 ± 0.0024 s	0.94
comptime/basics/2D sum (optimize = :before_enzyme)	0.0341 ± 0.0028 s	0.0334 ± 0.002 s	1.02
comptime/basics/2D sum (optimize = :only_enzyme)	27.8 ± 2 ms	29.5 ± 2.4 ms	0.943
comptime/basics/cos.(x) (optimize = :after_enzyme)	0.0393 ± 0.0021 s	0.0395 ± 0.0021 s	0.997
comptime/basics/cos.(x) (optimize = :all)	0.0378 ± 0.0017 s	0.0402 ± 0.002 s	0.939
comptime/basics/cos.(x) (optimize = :before_enzyme)	0.0376 ± 0.002 s	0.0395 ± 0.002 s	0.951
comptime/basics/cos.(x) (optimize = :only_enzyme)	0.0362 ± 0.0017 s	0.0367 ± 0.0019 s	0.987
comptime/basics/∇cos (optimize = :all)	0.061 ± 0.0037 s	0.0653 ± 0.0048 s	0.934
runtime/NN/ViT base (optimize = :after_enzyme)	6.44 s	6.65 s	0.968
runtime/NN/ViT base (optimize = :all)	6.62 s	6.58 s	1.01
runtime/NN/ViT base (optimize = :before_enzyme)	6.6 s	6.49 s	1.02
runtime/NN/ViT base (optimize = :only_enzyme)	7.85 s	7.94 s	0.989
runtime/NN/ViT tiny (optimize = :after_enzyme)	1.77 s	1.73 s	1.02
runtime/NN/ViT tiny (optimize = :all)	1.76 s	1.77 s	0.99
runtime/NN/ViT tiny (optimize = :before_enzyme)	1.81 s	1.78 s	1.02
runtime/NN/ViT tiny (optimize = :only_enzyme)	2.9 s	2.98 s	0.974
runtime/NN/vgg11 bn=false (optimize = :after_enzyme)	2.14 s	2.2 s	0.975
runtime/NN/vgg11 bn=false (optimize = :all)	2.16 s	2.2 s	0.985
runtime/NN/vgg11 bn=false (optimize = :before_enzyme)	2.16 s	2.15 s	1
runtime/NN/vgg11 bn=false (optimize = :only_enzyme)	2.02 s	1.94 s	1.04
runtime/NN/vgg11 bn=true (optimize = :after_enzyme)	2.31 s	2.29 s	1.01
runtime/NN/vgg11 bn=true (optimize = :all)	2.31 s	2.32 s	0.995
runtime/NN/vgg11 bn=true (optimize = :before_enzyme)	2.38 s	2.27 s	1.05
runtime/NN/vgg11 bn=true (optimize = :only_enzyme)	2.4 s	2.39 s	1
runtime/NN/vgg13 bn=false (optimize = :after_enzyme)	3.11 s	3.09 s	1.01
runtime/NN/vgg13 bn=false (optimize = :all)	2.77 s	2.94 s	0.941
runtime/NN/vgg13 bn=false (optimize = :before_enzyme)	3.06 s	3.05 s	1
runtime/NN/vgg13 bn=false (optimize = :only_enzyme)	2.9 s	2.84 s	1.02
runtime/NN/vgg13 bn=true (optimize = :after_enzyme)	3.19 s	3.26 s	0.979
runtime/NN/vgg13 bn=true (optimize = :all)	3.24 s	3.28 s	0.99
runtime/NN/vgg13 bn=true (optimize = :before_enzyme)	3.37 s	3.27 s	1.03
runtime/NN/vgg13 bn=true (optimize = :only_enzyme)	3.47 s	3.4 s	1.02
runtime/NN/vgg16 bn=false (optimize = :after_enzyme)	3.83 s	3.77 s	1.02
runtime/NN/vgg16 bn=false (optimize = :all)	3.68 s	3.61 s	1.02
runtime/NN/vgg16 bn=false (optimize = :before_enzyme)	3.88 s	3.78 s	1.03
runtime/NN/vgg16 bn=false (optimize = :only_enzyme)	3.61 s	3.6 s	1
runtime/NN/vgg16 bn=true (optimize = :after_enzyme)	4.28 s	4.15 s	1.03
runtime/NN/vgg16 bn=true (optimize = :all)	4.19 s	4.15 s	1.01
runtime/NN/vgg16 bn=true (optimize = :before_enzyme)	4.23 s	4.2 s	1.01
runtime/NN/vgg16 bn=true (optimize = :only_enzyme)	4.48 s	4.58 s	0.978
runtime/NN/vgg19 bn=false (optimize = :after_enzyme)	4.7 s	4.58 s	1.03
runtime/NN/vgg19 bn=false (optimize = :all)	4.56 s	4.57 s	0.999
runtime/NN/vgg19 bn=false (optimize = :before_enzyme)	4.6 s	4.59 s	1
runtime/NN/vgg19 bn=false (optimize = :only_enzyme)	4.54 s	4.47 s	1.02
runtime/NN/vgg19 bn=true (optimize = :after_enzyme)	5.11 s	5.2 s	0.981
runtime/NN/vgg19 bn=true (optimize = :all)	5.15 s	5.13 s	1
runtime/NN/vgg19 bn=true (optimize = :before_enzyme)	5.09 s	4.98 s	1.02
runtime/NN/vgg19 bn=true (optimize = :only_enzyme)	5.65 s	5.7 s	0.992
time_to_load	2.14 ± 0.06 s	2.27 ± 0.064 s	0.944

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).