prevent canonicalization where arguments have metadata

[shashi]

#546

[github-actions]

Benchmark Results

	master	ba61640...	t[master]/t[ba61640...]
overhead/acrule/a+2	1.12 ± 0.23 μs	1.1 ± 0.2 μs	1.02
overhead/acrule/a+2+b	1.1 ± 0.2 μs	1.1 ± 0.22 μs	1
overhead/acrule/a+b	0.334 ± 0.0064 μs	0.337 ± 0.0072 μs	0.992
overhead/acrule/noop:Int	25.1 ± 0.1 ns	26 ± 0.2 ns	0.965
overhead/acrule/noop:Sym	0.0441 ± 0.0073 μs	0.0396 ± 0.0066 μs	1.11
overhead/rule/noop:Int	0.0424 ± 0.0032 μs	0.0408 ± 0.0027 μs	1.04
overhead/rule/noop:Sym	0.0554 ± 0.0019 μs	0.0536 ± 0.0019 μs	1.03
overhead/rule/noop:Term	0.0557 ± 0.0017 μs	0.0536 ± 0.002 μs	1.04
overhead/ruleset/noop:Int	0.154 ± 0.0023 μs	0.156 ± 0.0022 μs	0.984
overhead/ruleset/noop:Sym	0.188 ± 0.0047 μs	0.183 ± 0.0057 μs	1.02
overhead/ruleset/noop:Term	4.56 ± 0.29 μs	4.31 ± 0.29 μs	1.06
overhead/simplify/noop:Int	0.233 ± 0.01 μs	0.209 ± 0.0097 μs	1.11
overhead/simplify/noop:Sym	0.244 ± 0.0046 μs	0.222 ± 0.0042 μs	1.1
overhead/simplify/noop:Term	0.0414 ± 0.0021 ms	0.0413 ± 0.0022 ms	1
overhead/simplify/randterm (+, *):serial	0.149 ± 0.0027 s	0.144 ± 0.0012 s	1.03
overhead/simplify/randterm (+, *):thread	0.102 ± 0.013 s	0.1 ± 0.013 s	1.02
overhead/simplify/randterm (/, *):serial	0.242 ± 0.0065 ms	0.245 ± 0.0061 ms	0.99
overhead/simplify/randterm (/, *):thread	0.27 ± 0.0083 ms	0.272 ± 0.0075 ms	0.993
overhead/substitute/a	0.0752 ± 0.0014 ms	0.072 ± 0.0014 ms	1.04
overhead/substitute/a,b	0.0641 ± 0.0012 ms	0.063 ± 0.0012 ms	1.02
overhead/substitute/a,b,c	21.3 ± 0.7 μs	23.9 ± 0.8 μs	0.891
polyform/easy_iszero	0.0391 ± 0.0029 ms	0.0373 ± 0.0022 ms	1.05
polyform/isone	2.8 ± 0.1 ns	2.8 ± 0.1 ns	1
polyform/iszero	2.36 ± 0.075 ms	2.2 ± 0.035 ms	1.07
polyform/simplify_fractions	3.2 ± 0.16 ms	3.01 ± 0.051 ms	1.06
time_to_load	5.34 ± 0.016 s	5.31 ± 0.02 s	1.01

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).

[shashi]

Oh wow this means that nothing in Symbolics will be canonicalized because of the VariableName metadata we attach in @variables -- @YingboMa are we able to get rid of that yet?

[shashi]

Made it less conservative, it should continue to work 👍

[YingboMa]

Unfortunately we still need the variable metadata.