keithreid-sfw / stern_brocot_comparison_julia_vs_python

speed comparison of Stern Brocot between Julia and Python

Stern Brocot comparison Julia vs Python

#Python version

Python 3.7.6 (default, Jan 8 2020, 19:59:22)

[GCC 7.3.0] :: Anaconda, Inc. on linux

#Julia Version

1.8.1 (2022-09-06)

Date 19 Feb 2023

#Box Spec

OS: Ubuntu 22.04.1 LTS x86_64

Kernel: 5.15.0-60-generic

Shell: bash 5.1.16

CPU: AMD Ryzen 9 3900X (24) @ 3.800G

Memory: 64Gb RAM

#Intent:

Implement Stern Brocot in Python and Julia using TDD and nothing too clever

For speed comparison

Roughly simlar logic in the two versions

I love Python and Julia This is a fair but real-life test in a nerd's attic Using the versions I had to hand And the way I normally code And the same archiecture on the same afternoon etc.

Python

1 layers 0.000002879590399970766 seconds

2 layers 0.000005017571400003362 seconds

4 layers 0.000014577579199976754 seconds

8 layers 0.00019980101569999533 seconds

12 layers 0.0033494253342999402 seconds

13 layers 0.0067446708239000145 seconds

14 layers 0.013931729978600016 seconds

15 layers 0.030198445934199937 seconds

16 layers withdraws, estimated to take 10 minutes total

Julia

1 layers elapsed time (ns): 1590 0.000002 seconds

2 layers elapsed time (ns): 1770 0.000002 seconds

4 layers elapsed time (ns): 1350 0.000001 seconds

8 layers elapsed time (ns): 1410 0.000001 seconds

12 layers elapsed time (ns): 1980 0.000002 seconds

13 layers elapsed time (ns): 1670 0.000002 seconds

14 layers elapsed time (ns): 2030 0.000002 seconds

15 layers elapsed time (ns): 2150 0.000002 seconds

16 layers elapsed time (ns): 2220 0.000002 seconds

24 layers elapsed time (ns): 2310 0.000002 seconds

32 layers blows up RAM