DEV Community

OpenSource for Webcrumbs

Posted on

Unlocked a new GPT functionality: performance calculations

So I was curious if using a super long hash would impact my apps performance.

Asked ChatGPT 4o:

1. What's the performance comparison between x and y?

In this example, x and y were:

const hash = crypto.createHash('sha256').update(message + code + (files ? files.map(f => f.buffer.toString('hex')).join('') : '') + uniqueValue).digest('hex');
Enter fullscreen mode Exit fullscreen mode
const hash = crypto.createHash('sha256').update(uniqueValue).digest('hex');
Enter fullscreen mode Exit fullscreen mode

Okay. It blabed a little bit. It even wrote the code so I could verify it. But I wanted an answer. So second and last iteration:

2. Can you run and let me know? Use fake data as input

And bum! Got my answer in a few seconds.

The performance comparison between the two methods is as follows:

Long Input Hash:
0.1235 seconds for 100,000 iterations

Short Input Hash:
0.1061 seconds for 100,000 iterations

As expected, the short input hash is slightly faster. The difference, while measurable, is relatively small. This indicates that if you are hashing a relatively small number of times, the performance impact will be minimal. However, if you need to optimize for the fastest possible hashing in a high-frequency context, using a short input value like a unique counter or timestamp is the better option. ​

Hope it helps!
Happy coding.

Top comments (0)