DAI 2023-2024 - Java IOs benchmarking - Jeremiah Steiner

[BetterBlood]

commit

Type	Opération	Buffer	Temps (ms)	Temps (ns)
Writing	binary_1B.bin	non-buffered	1	1480300
Reading	binary_1B.bin	non-buffered	0	389600
Writing	binary_1B.bin	buffered	1	1102100
Reading	binary_1B.bin	buffered	0	432300
Writing	text_1B.txt	buffered	1	1923200
Reading	text_1B.txt	buffered	0	660100
Writing	text_1B.txt	non-buffered	1	1279600
Reading	text_1B.txt	non-buffered	0	600600
Writing	binary_1KiB.bin	non-buffered	21	21197800
Reading	binary_1KiB.bin	non-buffered	9	9765700
Writing	binary_1KiB.bin	buffered	1	1158300
Reading	binary_1KiB.bin	buffered	1	1487900
Writing	text_1KiB.txt	buffered	2	2308200
Reading	text_1KiB.txt	buffered	1	1553600
Writing	text_1KiB.txt	non-buffered	8	8561700
Reading	text_1KiB.txt	non-buffered	4	4203900
Writing	binary_1MiB.bin	non-buffered	6918	6918317600
Reading	binary_1MiB.bin	non-buffered	6076	6076323000
Writing	binary_1MiB.bin	buffered	32	32630400
Reading	binary_1MiB.bin	buffered	47	47842000
Writing	text_1MiB.txt	buffered	78	78942500
Reading	text_1MiB.txt	buffered	110	110107800
Writing	text_1MiB.txt	non-buffered	203	203484500
Reading	text_1MiB.txt	non-buffered	173	173512000
Writing	binary_5MiB.bin	non-buffered	33086	33086780500
Reading	binary_5MiB.bin	non-buffered	30096	30096965000
Writing	binary_5MiB.bin	buffered	226	226607700
Reading	binary_5MiB.bin	buffered	226	226585800
Writing	text_5MiB.txt	buffered	334	334934100
Reading	text_5MiB.txt	buffered	382	382172900
Writing	text_5MiB.txt	non-buffered	577	577779600
Reading	text_5MiB.txt	non-buffered	440	440180400

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

size	type
1 B	not significant
1 KiB	buffered binary
1 MiB	buffered binary
5 MiB	buffered binary

Why is it more efficient than the other types of streams? cause it optimise the information transfer using buffers
What is the difference between binary data and text data? a char could take more than a byte
What is a character encoding? a meta information, let us know how to read bytes to understand them the right way
Why is this methodology important? to prevent loosing time on execution, to optimise our application

ChatGPT used for creating the tables using this prompt:
Récupère les données dans le texte ci-dessous, et formate certaines informations en tableau en format markdown. Les informations nécessaire sont : si le fichier est en binaire ou en texte, si l'utilisation d'un buffer est effectuée ou non, si c'est une lecture ou une écriture, le nombre de ms et le nombre de ns. Dans les données ci-dessous les infos sont groupés par deux lignes, je veux donc dans le tableau uns ligne pour chaque deux ligne du texte ci-dessous. Donne directement le .md
"Writing 1 bytes to binary_1B.bin with a non-buffered stream...
Elapsed time: 1ms / 1480300ns
{...}
Reading text_5MiB.txt with a non-buffered stream...
Elapsed time: 440ms / 440180400ns"

[ludelafo]

Great, thank you!

I guess the repository is private as I cannot access it. If you add us to your project or make it public, we can give you a feedback.

Good usage of ChatGPT.

[BetterBlood]

That's right, forget about this point, I've just added you two to my repo, I would appreciate feedback.

[ludelafo]

Thank you!

Everything seems fine by me. One thing you could improve is the usage of the try-with-resources type of try() { ... } as mentioned in the course or in the following tutorial: https://docs.oracle.com/javase/tutorial/essential/exceptions/tryResourceClose.html

The other thing you could improve is to add the .bin and .txt files to your gitignore files as these are generated by your application and take some space on your repository.

Other than this, everything looks good to me!

[BetterBlood]

That's done, thanks for the feedback.

[ludelafo]

You're welcome!