Binary Search Trees - Measuring Six Different Languages

November, 2015

I thought it would be fun to implement a binary search tree in six different languages, not only to show the implementation differences but to also show the memory/CPU performance results.

The results below are based off of a computer running Windows 10 Pro x64 with an Intel Core i7 4770k processor overclocked to 4.5 GHz and with 16 GB of RAM installed.

C

Generating random array with 2,000,000 values	Filling tree with 2,000,000 nodes	Traversing all 2,000,000 nodes	Peak memory used
22 ms	12.86 sec	128 ms	53 MB

C#

Version	Generating random array with 2,000,000 values	Filling tree with 2,000,000 nodes	Traversing all 2,000,000 nodes	Peak memory used
.NET 4.6/C# 4.0	25 ms	13.07 sec	128 ms	58 MB

Java (1.8)

Generating random array with 2,000,000 values	Filling tree with 2,000,000 nodes	Traversing all 2,000,000 nodes	Peak memory used
21 ms	12.2 sec	126 ms	70 MB

Go (1.5.1)

Generating random array with 2,000,000 values	Filling tree with 2,000,000 nodes	Traversing all 2,000,000 nodes	Peak memory used
66 ms	12.9 sec	130 ms	85 MB

Python

Version	Generating random array with 2,000,000 values	Filling tree with 2,000,000 nodes	Traversing all 2,000,000 nodes	Peak memory used
Python 2.7.10	1.9 sec	98 sec	750 ms	377 MB
Python 3.5.0	2.55 sec	114.23 sec	828 ms	220 MB

JavaScript

Version	Generating random array with 2,000,000 values	Filling tree with 2,000,000 nodes	Traversing all 2,000,000 nodes	Peak memory used
Node.js 5.1.0	34.64 ms	11.06 sec	89.38 ms	148 MB