DAI 2024-2025 - Java IOs benchmarking - Yoann Changanaqui

[Yoy017]

Link to the repo

Write

Size of file	Binary	Buf binary	Text	Buf text
1B	144	140	128	128
1KiB	160	140	134	133
1MiB	1191	232	262	163
5MiB	5191	252	345	250

(Time in ms)

Read

Size of file	Binary	Buf binary	Text	Buf text
1B	127	124	131	129
1KiB	290	127	227	132
1MiB	435	201	260	161
5MiB	1508	1496	427	248
(Time in ms)

Which type of stream is the most efficient for each use case?

The most efficient type is the Buffered.

Why is it more efficient than the other types of streams?

Because the buffer make the program use way less call of read() or write(). It improve the performance since the call of these method are few compare to the other types of streams.

What is the difference between binary data and text data?

A Text data file uses an character encoding that map binary code to a character. Meanwhile binary data file contains raw binary. It is not encoded and non-human readable. In general, binary data file are compiled programs.

What is a character encoding?

It is use to assign a binary code to a specific symbol. It's use for character and emojis. One of the most well-known is ASCII.

Why is this benchmark methodology important?

First to experiment writing IOs files that will be usefull for the Practical work 1 of this course. Secondly to see in what way one specific method (buffered) is the most efficient !

[ludelafo]

Thank you!

You can always find the solution in the solution repository if you ever need to compare your work with the official one here: https://github.com/heig-vd-dai-course/heig-vd-dai-course-solutions.

Do not hesitate to ask questions if something is not clear, either in this discussion or in the Q&A category of the GitHub Discussions here: https://github.com/orgs/heig-vd-dai-course/discussions/categories/q-a.