Re: Benchmarking PostgreSQL against others on Windows?

On Nov 27, 2006, at 5:24 AM, Tom Allison wrote:
> Basically it comes down to this:
> simple SELECT statements: mysql almost always wins.
> Everything else is mixed.
> There are a few cases where postgresql does really poorly but the
> operations are rare daily activity.
> INSERT/UPDATE/DELETE postgresql "generally" does better.
>
> The problem with the tests is that these are typically done with
> single users at a time and they don't exercise any of the other
> benefits that postgresql offers. But I would suggest you set up
> two versions of some aspect of your working environment: one in
> postgres and on in MySQL and run them for a few weeks using mock
> client scripts to perform different actions.

Actually, the folks over at http://tweakers.net recently benchmarked
MySQL and PostgreSQL with their website code. IIRC, this is mostly
"simple selects", yet PostgreSQL clearly outperformed MySQL once you
had a few concurrent connections happening.

My experience is that the only places you'll readily see MySQL
outperform PostgreSQL is when you're only using a single connection,
and probably not with complex queries.

Of course, the first rule of benchmarking is to use your own
application.
--
Jim Nasby jim(at)nasby(dot)net
EnterpriseDB http://enterprisedb.com 512.569.9461 (cell)