Home/ Questions/Q 8350463
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-06-09T08:19:52+00:00

If I want to compare the speed of 2 implementations of an algorithm, is

  • 0

If I want to compare the speed of 2 implementations of an algorithm, is there any gotcha about the JIT I should be aware of? For example, how many iterations does it take the JIT to “warm-up” (if it does – I am not very familiar with the JIT)? I have a merge-like algorithm in mind, where I want to compare using an iterator and using raw primitive arrays directly. I want to understand how well the JIT can handle the iterator-based algorithm.

My fear is that my comparison wouldn’t be relevant because I didn’t let the benchmark run long enough, and the JIT didn’t really have a chance to do its job.

Also, what would be the best reference to understand the JIT in depth (besides reading source code).

  • One more question: is -server the best option to use to tell the JIT to work hard?
Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    is there any gotcha about the JIT I should be aware of?

    too many to mention. 😉

    Don’t forget the CPU and cache does its own optimisation and performance improvements.

    , how many iterations does it take the JIT to “warm-up” (if it does – I am not very familiar with the JIT)?

    At least 10,000. I would perform the test for 2-10 seconds, a number of times.

    what would be the best reference to understand the JIT in depth

    I don’t think there is one “best” reference. Its a very complex subject.

    • 0