quicksort performance tests

[ViralBShah]

Currently, the quicksort performance tests across various languages are written in a variety of different ways. It would be best to implement the same algorithm, without randomness across all languages.

[StefanKarpinski]

Remaining todos here:

• Someone needs to go through all the different versions and make sure they're doing the same thing.
• Use a common simple pseudorandom number generator so that all languages are sorting the same array.

[ViralBShah]

Only the javascript version remains now. We should update our performance table on the website when the js version is done.

[ViralBShah]

Also, we can simply sort [n:-1:1] and avoid writing a pseudorandom number generator.

[StefanKarpinski]

That's a lousy test of qsort because it only tests one swapping pattern over and over instead of mixing it up.

[ViralBShah]

I think it is good enough for the purpose. Of course, an RNG is only 5 lines or so, but you have to write it in every language.

-viral

On 17-Mar-2012, at 11:57 PM, Stefan Karpinski wrote:

That's a lousy test of qsort because it only tests one swapping pattern over and over instead of mixing it up.

---

Reply to this email directly or view it on GitHub:
#562 (comment)

[Bluebugs]

Or just put the same huge randonmly filled array in each language. That would make it almost just a copy past.

[ViralBShah]

We can get the JS version of qsort from:

https://bugzilla.mozilla.org/show_bug.cgi?id=739016
https://groups.google.com/forum/?fromgroups#!topic/mozilla.dev.tech.js-engine/Ct6JsfJEjdI

[StefanKarpinski]

This seems to be in good enough shape now.