X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.144]) by svr1.postgresql.org (Postfix) with ESMTP id E3FB552967 for ; Thu, 21 Apr 2005 01:35:45 -0300 (ADT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (av.hub.org [200.46.204.144]) (amavisd-new, port 10024) with ESMTP id 53330-01 for ; Thu, 21 Apr 2005 04:35:36 +0000 (GMT) Received: from sss.pgh.pa.us (sss.pgh.pa.us [66.207.139.130]) by svr1.postgresql.org (Postfix) with ESMTP id 70DA353777 for ; Thu, 21 Apr 2005 01:35:34 -0300 (ADT) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.13.1/8.13.1) with ESMTP id j3L4ZWq5011217; Thu, 21 Apr 2005 00:35:32 -0400 (EDT) To: John A Meinel Cc: Joel Fradkin , Postgresql Performance Subject: Re: Joel's Performance Issues WAS : Opteron vs Xeon In-reply-to: <4266F7B3.7050004@arbash-meinel.com> References: <001901c54608$b70ccfc0$797ba8c0@jfradkin> <4266F7B3.7050004@arbash-meinel.com> Comments: In-reply-to John A Meinel message dated "Wed, 20 Apr 2005 19:45:39 -0500" Date: Thu, 21 Apr 2005 00:35:32 -0400 Message-ID: <11216.1114058132@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at hub.org X-Spam-Status: No, hits=0.007 tagged_above=0 required=5 tests=AWL X-Spam-Level: X-Archive-Number: 200504/548 X-Sequence-Number: 11945 John A Meinel writes: > Joel Fradkin wrote: >> Postgres was on the second run >> Total query runtime: 17109 ms. >> Data retrieval runtime: 72188 ms. >> 331640 rows retrieved. > How were you measuring "data retrieval time"? I suspect he's using pgadmin. We've seen reports before suggesting that pgadmin can be amazingly slow, eg here http://archives.postgresql.org/pgsql-performance/2004-10/msg00427.php where the *actual* data retrieval time as shown by EXPLAIN ANALYZE was under three seconds, but pgadmin claimed the query runtime was 22 sec and data retrieval runtime was 72 sec. I wouldn't be too surprised if that time was being spent formatting the data into a table for display inside pgadmin. It is a GUI after all, not a tool for pushing vast volumes of data around. It'd be interesting to check the runtimes for the same query with LIMIT 3000, ie, see if a tenth as much data takes a tenth as much processing time or not. The backend code should be pretty darn linear in this regard, but maybe pgadmin isn't. regards, tom lane