Interpreter benchmarks

All workloads

Workload	pipit median	Go	CPython	PyPy	tengo	scriggo	mvm	yaegi
Smoke: integer addition	465 µs	0.05×	3.31×	1.09×	6.50×	5.61×	2.16×	3.44×
Iterative Fibonacci	3.00 ms	0.04×	6.25×	3.43×	n/a	5.12×	6.31×	11.6×
Word frequency	131 ms	0.15×	2.20×	0.39×	7.07×	3.38×	4.80×	3.36×
Levenshtein, 1k pairs	74.9 ms	0.07×	2.32×	0.40×	8.31×	4.24×	6.33×	9.27×
Mini JSON parse	117 ms	0.12×	1.46×	1.03×	n/a	unsupported	2.98×	14.1×
Expression evaluator	1.21 s	0.05×	1.08×	0.21×	3.45×	unsupported	3.47×	8.12×
LRU cache	87.4 ms	0.18×	0.97×	0.20×	1.96×	unsupported	3.21×	8.64×
Dijkstra shortest paths	65.1 ms	0.08×	1.30×	0.65×	1.92×	2.31×	2.72×	3.51×
Sudoku solver, 100 boards	18.4 ms	0.06×	1.42×	1.41×	2.90×	unsupported	2.16×	6.19×
Game of Life	1.78 s	0.05×	3.34×	0.09×	5.72×	3.87×	5.05×	5.77×
Markov text generation	23.0 ms	0.23×	1.19×	0.49×	2.23×	2.19×	2.50×	1.87×
Trie of 50k words	279 ms	0.03×	1.89×	0.19×	4.69×	unsupported	3.65×	4.17×
Brainfuck interpreter	163 µs	0.17×	2.20×	15.7×	73.3×	3.30×	10.7×	10.5×
Polymorphic AST eval	31.9 ms	0.06×	1.63×	2.70×	3.73×	unsupported	7.52×	34.5×
Mandelbrot (FP)	24.5 ms	0.14×	5.98×	0.54×	9.34×	8.21×	7.75×	8.83×
Open-addressing hash map	292 ms	0.08×	2.25×	0.32×	5.46×	unsupported	2.61×	5.01×
Parallel word count (Monte Cristo)	90.4 ms	0.25×	2.10×	3.30×	n/a	unsupported	2.83×	n/a
Closures pipeline	80.4 ms	0.03×	1.44×	0.31×	2.25×	1.74×	1.85×	8.16×
Generic numeric pipeline	17.7 ms	0.04×	1.83×	0.58×	4.00×	unsupported	2.26×	2.37×
Type switches	50.3 ms	0.01×	2.49×	0.57×	2.49×	2.21×	3.01×	4.75×
Invert binary tree	61.2 ms	0.10×	1.31×	0.70×	2.94×	n/a	2.70×	9.37×
Dense neural-network layer	831 µs	0.18×	30.2×	7.47×	27.5×	17.0×	21.9×	19.7×
N-body solar system	207 ms	0.04×	2.02×	0.13×	5.26×	4.54×	4.60×	5.63×
Per-line SHA-256 (FFI overhead)	31.7 ms	0.06×	1.51×	0.71×	2.52×	2.57×	2.64×	1.76×

Interpreter benchmarks live data