Message-ID: <4463BF89.6050702@joeconway.com>
Date: Thu, 11 May 2006 15:49:45 -0700
From: Joe Conway <mail@joeconway.com>
User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US;
	rv:1.7.13) Gecko/20060501 Fedora/1.7.13-1.1.fc5
MIME-Version: 1.0
To: "Jim C. Nasby" <jnasby@pervasive.com>
CC: Adam Palmblad <apalmblad@dataunison.com>, pgsql-performance@postgresql.org
Subject: Re: Dynamically loaded C function performance
References: <6B5AF6293A289F45826220B17ABE79370D5ED7@BORON.aers.local>
	<20060511220523.GT99570@pervasive.com>
In-Reply-To: <20060511220523.GT99570@pervasive.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

Jim C. Nasby wrote:
> On Fri, May 05, 2006 at 03:47:53PM -0700, Adam Palmblad wrote:
> 
>>Hi,
>>We've got a C function that we use here and we find that for every
>>connection, the first run of the function is much slower than any
>>subsequent runs.  ( 50ms compared to 8ms)
>>
>>Besides using connection pooling, are there any options to improve
>>performance?
> 
> In my experience, connection startup takes a heck of a lot longer than
> 50ms, so why are you worrying about 50ms for the first run of a
> function?
> 
> BTW, sorry, but I don't know a way to speed this up, either.

I think Tom nailed the solution already in a nearby reply -- see 
preload_libraries on this page:

http://www.postgresql.org/docs/8.1/interactive/runtime-config-resource.html

Joe