[TYPO3-dev] Quick speed test mysql vs mysqli, for now a win for mysql

[Steffen Müller]

Hi.

Martin Kutschker schrieb:
> 
> I think that the cached version is faster because a) the query plan is
> not expensive for the simple query and b) the query cache is large
> enough to hold all of those random queries.

Regarding a TYPO3 system in production:
I have an query cache hit:insert rate of 10:1
Means: an average of ~91% of all queries are retrieved from the cache.

mysql> SHOW STATUS LIKE '%Qcache%ts';
+----------------+-----------+
| Variable_name  | Value     |
+----------------+-----------+
| Qcache_hits    | 300657676 |
| Qcache_inserts | 30671648  |
+----------------+-----------+

Prepared stements or not: query cache is hard to beat, if disk access is
the bottleneck.

>From MySQL 5.1.17, query cache partly supports prepared statements, some
more support since 5.1.21:

http://dev.mysql.com/doc/refman/5.1/en/query-cache-how.html

@Ries: can you please repeat the test with one of these MySQL versions?

-- 
cheers,
Steffen