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 x64 octa-core

 spectral-norm benchmark N=550

Each chart bar shows how many times slower, one ↓ spectral-norm program was, compared to the fastest program.

These are not the only programs that could be written. These are not the only compilers and interpreters. These are not the only programming languages.

Column × shows how many times more each program used compared to the benchmark program that used least.

		sort		sort	sort
×	Program Source Code	CPU secs	Elapsed secs	Memory KB	Code B	≈ CPU Load
1.0	PyPy 3 #8	0.09	0.09	?	594	0% 18% 100% 0% 11% 0% 0% 11%
1.0	PyPy 2 #8	0.09	0.09	?	594	0% 0% 0% 100% 18% 0% 0% 0%
1.1	Python development version #2	0.10	0.10	?	394	0% 0% 100% 0% 0% 0% 10% 0%
1.2	Nuitka #2	0.10	0.11	?	394	8% 0% 91% 0% 0% 0% 0% 0%
1.3	Python 3 #2	0.11	0.11	?	394	100% 0% 0% 0% 0% 0% 8% 0%
1.9	PyPy 2 #6	0.17	0.17	?	498	0% 100% 0% 19% 6% 12% 0% 0%
2.0	PyPy 3 #6	0.18	0.18	?	498	0% 100% 15% 0% 0% 0% 6% 0%
3.6	Python development version #3	0.31	0.32	2,560	642	6% 0% 0% 100% 0% 6% 3% 0%
3.7	Python 3 #3	0.32	0.33	1,792	642	100% 3% 0% 0% 0% 0% 3% 0%
3.9	Nuitka #3	0.31	0.35	3,456	642	100% 0% 0% 0% 0% 0% 0% 6%
4.4	Pyston #5	1.40	0.39	43,572	595	55% 68% 57% 37% 37% 37% 37% 39%
6.7	Python development version #5	2.21	0.60	56,516	575	41% 46% 33% 48% 54% 67% 53% 42%
7.3	Python 3 #5	2.48	0.65	55,220	575	59% 50% 20% 38% 59% 56% 55% 55%
7.8	Nuitka #5	2.61	0.70	58,040	575	46% 53% 41% 30% 38% 67% 55% 64%
7.8	PyPy 2 #5	1.12	0.70	333,172	595	32% 30% 46% 26% 10% 6% 13% 8%
11	Graal #8	1.97	0.96	729,476	594	26% 26% 57% 35% 19% 12% 1% 36%
13	Python 2 #5	4.23	1.13	11,728	595	78% 71% 51% 30% 25% 35% 40% 59%
15	Pyston #6	1.30	1.30	8,048	498	0% 99% 2% 0% 0% 0% 1% 2%
15	PyPy 3 #5	2.24	1.30	90,768	575	48% 29% 23% 24% 18% 16% 14% 16%
15	Numba #2	1.31	1.34	131,484	416	15% 9% 10% 8% 8% 9% 13% 99%
16	Numba	1.39	1.41	130,424	667	21% 13% 15% 17% 14% 18% 99% 14%
16	Pyston #8	1.43	1.43	8,160	594	3% 1% 0% 1% 1% 1% 1% 100%
24	Nuitka #6	2.11	2.11	11,264	498	100% 0% 0% 0% 2% 0% 0% 0%
24	Nuitka #8	2.12	2.13	11,392	594	0% 0% 1% 0% 0% 0% 0% 100%
24	Python development version #8	2.15	2.15	9,364	594	0% 0% 0% 0% 1% 0% 2% 100%
24	Python 3 #6	2.15	2.16	10,384	498	100% 1% 0% 0% 0% 0% 0% 0%
24	Python development version #6	2.16	2.17	8,872	498	0% 0% 100% 0% 2% 0% 0% 1%
25	Python 3 #8	2.21	2.21	10,452	594	0% 0% 0% 100% 0% 0% 0% 1%
27	Python 2 #6	2.36	2.36	6,640	498	0% 0% 0% 0% 1% 100% 0% 0%
30	Graal #6	8.10	2.65	921,616	498	8% 0% 65% 70% 69% 1% 96% 2%
30	Python 2 #8	2.70	2.71	6,592	594	0% 0% 0% 100% 0% 2% 1% 0%
63	Jython #8	8.79	5.59	3,540	594	23% 31% 26% 41% 10% 8% 8% 11%
70	MicroPython #6	6.21	6.22	4,200	498	0% 100% 0% 0% 0% 0% 0% 1%
83	Jython #6	10.41	7.38	3,660	498	10% 10% 9% 8% 58% 6% 26% 17%
612	RustPython #6	54.47	54.49	15,896	498	0% 1% 1% 1% 0% 1% 100% 1%
751	RustPython #8	66.80	66.81	15,136	594	1% 100% 1% 0% 0% 1% 0% 0%
missing benchmark programs
	IronPython	No program
	Cython	No program
	Shedskin	No program
	Grumpy	No program

 spectral-norm benchmark : Eigenvalue using the power method

diff program output N = 100 with this output file to check your program is correct before contributing.

We are trying to show the performance of various programming language implementations - so we ask that contributed programs not only give the correct result, but also use the same algorithm to calculate that result.

Each program should calculate the spectral norm of an infinite matrix A, with entries a₁₁=1, a₁₂=1/2, a₂₁=1/3, a₁₃=1/4, a₂₂=1/5, a₃₁=1/6, etc

For more information see challenge #3 in Eric W. Weisstein, "Hundred-Dollar, Hundred-Digit Challenge Problems" and "Spectral Norm".

From MathWorld--A Wolfram Web Resource.
http://mathworld.wolfram.com/Hundred-DollarHundred-DigitChallengeProblems.html
http://mathworld.wolfram.com/SpectralNorm.html

Thanks to Sebastien Loisel for this benchmark.

  Home   Conclusions   License   Play