Date: Mon, 21 May 2007 23:39:39 +0200
To: "Jim C. Nasby" <decibel@decibel.org>
Subject: Re: Postgres Benchmark Results
From: PFC <lists@peufeu.com>
Cc: pgsql-performance@postgresql.org
Content-Type: text/plain; format=flowed; delsp=yes; charset=utf-8
MIME-Version: 1.0
References: <op.tsmrv71ecigqcu@apollo13> <20070521210522.GN62346@nasby.net>
Content-Transfer-Encoding: 7bit
Message-ID: <op.tso44doqcigqcu@apollo13>
In-Reply-To: <20070521210522.GN62346@nasby.net>
User-Agent: Opera Mail/9.10 (Linux)

On Mon, 21 May 2007 23:05:22 +0200, Jim C. Nasby <decibel@decibel.org>  
wrote:

> On Sun, May 20, 2007 at 04:58:45PM +0200, PFC wrote:
>>
>> 	I felt the world needed a new benchmark ;)
>> 	So : Forum style benchmark with simulation of many users posting and
>> viewing forums and topics on a PHP website.
>>
>> 	http://home.peufeu.com/ftsbench/forum1.png
>
> Any chance of publishing your benchmark code so others can do testing?
> It sounds like a useful, well-thought-out benchmark (even if it is
> rather specialized).

	Yes, that was the intent from the start.
	It is specialized, because forums are one of the famous server killers.  
This is mostly due to bad database design, bad PHP skills, and the  
horrendous MySQL FULLTEXT.
	I'll have to clean up the code and document it for public consumption,  
though.
	However, the Python client is too slow. It saturates at about 1000 hits/s  
on a Athlon 64 3000+, so you can forget about benchmarking anything meaner  
than a Core 2 duo.

> Also, I think it's important for you to track how long it takes to
> respond to requests, both average and maximum. In a web application no
> one's going to care if you're doing 1000TPS if it means that every time
> you click on something it takes 15 seconds to get the next page back.
> With network round-trip times and what-not considered I'd say you don't
> want it to take any more than 200-500ms between when a request hits a
> webserver and when the last bit of data has gone back to the client.

	Yeah, I will do that too.

> I'm guessing that there's about 600MB of memory available for disk
> caching? (Well, 600MB minus whatever shared_buffers is set to).

	It's about that. The machine has 1 GB of RAM.