I read here and there that using the utf8_unicode_ci collation ensures a better treatment of unicode text (for example, it knowns how to expand characters such as 'œ' into 'oe' for searching and ordering) compared to the default utf8_general_ci which basically just strips diacritics. Unfortunately, both sources indicate that utf8_unicode_ci is slightly slower than utf8_general_ci.

So my question is: what does "slightly slower" mean? Has anyone run benchmarks? Are we talking about a -0.01% performance impact or rather something like -25%?

Thanks for your help.

link|improve this question

17% accept rate
As far as a benchmark, why not use the query time? I may be an idiot, but what if you ran up a VM and test the query time on a large complicated query for both character encodings? (I have not seen benchmarking done for this before) – Ablue Dec 30 '10 at 13:00
feedback

1 Answer

I didn't see any benchmark, but you can run your own using the BENCHMARK function:

BENCHMARK(count,expr)

As advised by Matthew you may run a parallel installation of MYSQL, but consider that there could be a huge difference between different architecture (sparc, intel, 32bit, 64bit, ...).

link|improve this answer
feedback

Your Answer

 
or
required, but never shown

Not the answer you're looking for? Browse other questions tagged or ask your own question.