Re: Performance with new nested-xacts code - Mailing list pgsql-hackers

From Alvaro Herrera
Subject Re: Performance with new nested-xacts code
Date
Msg-id 20040701135206.GA12738@dcc.uchile.cl
Whole thread Raw
In response to Re: Performance with new nested-xacts code  (Alvaro Herrera <alvherre@dcc.uchile.cl>)
List pgsql-hackers
On Thu, Jul 01, 2004 at 09:07:11AM -0400, Alvaro Herrera wrote:

> > Were your numbers also taken with --enable-cassert?  It might be
> > instructive to compare numbers taken without.
> 
> Ah, yes, it was with asserts enabled.  I'll try again.

With asserts disabled the situations seems reverted:
   Yesterday   Today   325.76      277.92   304.12      272.92   264.32      463.23   282.75      252.96   284.92
268.38  327.97      244.47   364.94      243.23
 
Min 264.32      243.23
Max 364.94      463.23
Avg 307.82      289.02

There's a lot more variance though, which I can easily attribute to the
server being more loaded (this is a Uni server; last night there was
nobody connected and today there's quite a few people, though the
majority is just using tin, pine or elm).

(These numbers are not comparable to what I posted yesterday because I
changed the checkpoint_segments parameter in between.)

-- 
Alvaro Herrera (<alvherre[a]dcc.uchile.cl>)
"Escucha y olvidarás; ve y recordarás; haz y entenderás" (Confucio)



pgsql-hackers by date:

Previous
From: Dennis Bjorklund
Date:
Subject: Re: Bug with view definitions?
Next
From: Tom Lane
Date:
Subject: nested-xacts cursors (was Re: Performance with new nested-xacts code)