Re: benchmarking postgres - Mailing list pgsql-hackers

From Ross J. Reedstrom
Subject Re: benchmarking postgres
Date
Msg-id 20020213162300.GB9356@rice.edu
Whole thread Raw
In response to Re: benchmarking postgres  (mlw <markw@mohawksoft.com>)
List pgsql-hackers
This comes up about once every 6 months. Please take it off HACKERS to
ADVOCACY (do we have such a thing?), or some such. Various members of
the PostgreSQL community have tried to work with the MySQL people in
the past to address 'issues' with their 'benchmark': it never works out.

Ross

On Wed, Feb 13, 2002 at 08:41:50AM -0500, mlw wrote:
> Ola Sundell wrote:
> > 
> > On Wed, 13 Feb 2002, mlw wrote:
> > 
> > > Christopher Kings-Lynne wrote:
> > > >
> > > > Look at this: (top one)
> > > >
> > > > http://www.mysql.com/information/benchmarks.html
> > > >
> > > > Does anyone feel like running the MySQL benchmark against postgres 7.2 to
> > > > see if there's been a real speed improvement??
> > > >
> > > > Chris
> > >
> > > These guys are just A$%$%holes. We have to come up with a benchmark which shows
> > > the the difference between a stupid little file-locking single user toy, and a
> > > real tansactional system.
> > >
> > > Maybe we too can put in little snide remarks about MySQL.
> > 
> > Now, let's be a bit sensible, here. MySQL is a great product, if you want
> > a single-user SQL interface to flat files. It is blazingly fast when it
> > comes to retrieving information in an environment where there is little or
> > no data change.
> 
> The snide remarks on the page about things not working was a bit much. I was
> ticked off. On a more serious note, MySQL isn't even really SQL. It supports a
> lot of the syntax, but none of the intentions. Things like sub-selects are
> vital to being able to model a problem. Transactions are vital to predictable
> behavior. High concurrency is vital to "real" performance.
> 
> I have said it at least a hundred times before, I have never been able to
> finish a project started in MySQL. I always come across something that the
> database *must* do, but MySQL can't.
> 
> It is clear that anyone who runs a single user benchmark against a database
> server capable of multiple connections is not testing their system in its
> intended mode of use. They are resorting to the worst sort of microsoftian
> benchmark FUD.
> 
> > 
> > We all know the strenghts of postgresql. It is a fully-featured
> > transactional database. MySQL is not, but it is neither stupid, nor a
> > toy. It has its purposes, as does postgresql.
> 
> What purpose does MySQL fit? It isn't very good at doing the sorts of things
> SQL is supposed to do and there are faster database libraries (ala Berkeley
> DB). What would be the point of using MySQL for anything?
> 
> ---------------------------(end of broadcast)---------------------------
> TIP 5: Have you checked our extensive FAQ?
> 
> http://www.postgresql.org/users-lounge/docs/faq.html


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Odd statistics behaviour in 7.2
Next
From: Peter Eisentraut
Date:
Subject: Re: benchmarking postgres