Expression evaluator

Tokenise then evaluate 10,000 arithmetic expressions (mixed +, -, *, /, parens).

Runtime · median per inner-loop window

median of 10 runs

Native Gocompiled

76.5 ms0.04×

Piko interpbytecode VM

2.15 sbaseline

CPython 3.13bytecode VM

1.43 s0.66×

PyPy 7.3tracing JIT

221 ms0.10×

Ttengobytecode VM

4.36 s2.02×

Sscriggobytecode VM

unsupported in scriggo

n/a

Mmvmbytecode VM

5.21 s2.42×

YyaegiAST walker

12.8 s5.93×

Runner	N	Compile	Runtime	P95	Stddev	RSS	vs piko	Status
Native Gocompiled	10	180 ms	76.5 ms	78.1 ms	1.51 ms	145 MiB	0.04×	OK
Piko interpbytecode VM	10	2.46 ms	2.15 s	2.35 s	65.0 ms	2.88 GiB	1.00×	OK
CPython 3.13bytecode VM	10	572 µs	1.43 s	1.46 s	26.2 ms	n/a	0.66×	OK
PyPy 7.3tracing JIT	10	467 µs	221 ms	258 ms	11.6 ms	n/a	0.10×	OK
tengobytecode VM	10	387 µs	4.36 s	4.42 s	47.6 ms	2.19 GiB	2.02×	OK
scriggobytecode VM	0	n/a	n/a	n/a	n/a	n/a	n/a	unsupported
mvmbytecode VM	10	752 µs	5.21 s	5.32 s	71.2 ms	66 MiB	2.42×	OK
yaegiAST walker	10	730 µs	12.8 s	13.1 s	190 ms	67 MiB	5.93×	OK

Workload & symmetry rules

Workload

For each of 10,000 randomly-generated infix expressions: tokenise with a byte-loop lexer, parse via shunting-yard, evaluate over an integer stack.

Source code