Python Interpreters Benchmarks
x64 ArchLinux : AMD® Ryzen 7 4700U®

- all benchmarks - n-body fannkuch-redux meteor-contest fasta fasta-redux fibonacci fib50 spectral-norm reverse-complement mandelbrot k-nucleotide regex-dna pidigits chameneos-redux templates thread-ring binary-trees-redux binary-trees pystone richards iobench jsonbench - all languages - Python 2 PyPy Python 3 Python dev PyPy 3 Jython IronPython Cython Nuitka Shedskin Numba Pyston MicroPython Grumpy Graal RustPython

vs Python 2 PyPy Python 3 Python dev PyPy 3 Jython IronPython Cython Nuitka Shedskin Numba Pyston MicroPython Grumpy Graal RustPython x64 octa-core

 1 : Are the Numba programs faster? At a glance.

Each chart bar shows, for one unidentified benchmark, how much the fastest Numba program used compared to the fastest PyPy program.

(Memory use is only compared for tasks that require memory to be allocated.)

These are not the only compilers and interpreters. These are not the only programs that could be written. These are not the only tasks that could be solved. These are just 10 tiny examples.

 2 : Are the Numba programs faster? Approximately.

Each table row shows, for one named benchmark, how much the fastest Numba program used compared to the fastest PyPy program.

(Memory use is only compared for tasks that require memory to be allocated.)

Numba used what fraction? used how many times more?
Benchmark	Time	Memory	Code
iobench †	1/3		±
templates †	±		±
fib50 †	3×		±
fibonacci	9×	?	±
n-body †	13×		±
spectral-norm †	15×	?	±
binary-trees	61×	2×	±

Numba used what fraction? used how many times more?
Time-used	|-	|---	25%	median	75%	---|	-|
(Elapsed secs)	1/3	1/3	2×	9×	14×	32×	61×

† possible mismatch - one-core program compared to multi-core program.

± read the measurements and then read the program source code.

 3 : Are the Numba programs faster? Measurements.

These are not the only tasks that could be solved. These are just 10 tiny examples. These are not the only compilers and interpreters. These are not the only programs that could be written.

For each named benchmark, measurements of the fastest Numba program are shown for comparison against measurements of the fastest PyPy program.

Program Source Code	CPU secs	Elapsed secs	Memory KB	Code B	≈ CPU Load
iobench
Numba	4.86	4.88	131,784	400	37% 7% 7% 9% 61% 5% 16% 4% †
PyPy	14.73	14.73	80,864	370	0% 0% 100% 0% 0% 0% 1% 0% †
templates
Numba	1.62	1.64	121,232	344	24% 19% 99% 23% 24% 24% 26% 19% †
PyPy	2.37	2.41	98,384	307	0% 6% 94% 0% 0% 0% 1% 0% †
fib50
Numba	1.47	1.48	119,832	165	15% 14% 8% 4% 5% 13% 10% 99% †
PyPy	0.48	0.48	77,576	130	4% 100% 0% 2% 2% 0% 2% 0% †
fibonacci
Numba	1.57	1.59	120,220	164	38% 28% 31% 98% 41% 51% 31% 52%
PyPy	0.17	0.17	?	181	6% 0% 0% 0% 0% 6% 11% 100%
n-body
Numba	42.59	42.63	148,116	1347	30% 67% 10% 9% 37% 16% 10% 15% †
PyPy	3.29	3.29	74,856	1337	0% 100% 1% 0% 1% 1% 1% 1% †
spectral-norm
Numba	1.31	1.34	131,484	416	15% 9% 10% 8% 8% 9% 13% 99% †
PyPy	0.09	0.09	?	594	0% 0% 0% 100% 18% 0% 0% 0% †
binary-trees
Numba	357.88	63.59	154,632	734	73% 70% 76% 94% 76% 74% 69% 75%
PyPy	2.27	1.04	91,028	743	32% 67% 25% 28% 24% 28% 22% 24%
richards
	No program
PyPy	0.18	0.18	?	2423	18% 5% 100% 0% 0% 5% 5% 0%
mandelbrot
Numba	61.51	8.68	133,960	3886	94% 91% 96% 93% 93% 98% 92% 95%
	No program
regex-dna
	No program
PyPy	0.10	0.11	?	501	100% 0% 0% 0% 0% 0% 0% 0%
reverse-complement
	No program
PyPy	0.06	0.06	?	432	0% 14% 0% 0% 0% 14% 14% 100%
pidigits
	No program
PyPy	2.98	2.99	81,732	322	100% 0% 0% 0% 1% 0% 0% 0%
pystone
	No program
PyPy	0.08	0.09	?	2301	100% 10% 0% 0% 0% 20% 0% 11%
fasta
	No program
PyPy	1.75	1.75	77,640	900	2% 2% 2% 2% 100% 3% 2% 1%
fannkuch-redux
	No program
PyPy	1.68	0.58	322,008	1009	41% 19% 66% 16% 43% 68% 66% 32%
fasta-redux
	No program
PyPy	0.19	0.19	?	1115	0% 0% 5% 100% 5% 5% 0% 0%
k-nucleotide
	No program
PyPy	0.09	0.09	?	593	22% 18% 0% 100% 0% 0% 0% 0%
jsonbench
	No program
PyPy	0.90	0.90	80,556	322	0% 0% 100% 3% 1% 0% 1% 0%
thread-ring
	No program
PyPy	0.11	0.11	?	407	0% 8% 92% 0% 9% 0% 0% 8%
meteor-contest
	No program
PyPy	0.67	0.67	80,972	1485	4% 0% 1% 99% 0% 4% 0% 0%
chameneos-redux
	No programs
binary-trees-redux
	No programs

† possible mismatch - one-core program compared to multi-core program.

 4 : Are there other Numba programs for these benchmarks?

Remember - those are just the fastest Numba and PyPy programs measured on this OS/machine. Check if there are other implementations of these benchmark programs for Numba.

Python 2 PyPy Python 3 Python dev PyPy 3 Jython IronPython Cython Nuitka Shedskin Numba Pyston MicroPython Grumpy Graal RustPython

Maybe one of those other Numba programs is fastest on a different OS/machine.

 5 : Are there other faster programs for these benchmarks?

Remember - those are just the fastest Numba and PyPy programs measured on this OS/machine. Check if there are faster implementations of these benchmark programs for other programming languages.

n-body fannkuch-redux meteor-contest fasta fasta-redux fibonacci fib50 spectral-norm reverse-complement mandelbrot k-nucleotide regex-dna pidigits chameneos-redux templates thread-ring binary-trees-redux binary-trees pystone richards iobench jsonbench

Maybe one of those other programs is fastest on a different OS/machine.

 Numba : Numba JIT compiler 

  Home   Conclusions   License   Play