DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma;
	h=message-id:date:from:to:subject:cc:in-reply-to:mime-version
	:content-type:content-transfer-encoding:content-disposition
	:references;
	b=Kl6QYRuq0xiN3HTCOzv2S+6r8r36kfLbfaMOQf+4Z0XcOlp9NY10VtBaTCo4Lf8G7j
	9FqsQE8pPLSLC68ckjoKplhGuPmhrWuua7jq0FxtcZv/bc75Wu901PotdhlUYPa0MP12
	07Qrfs1bPkrb6o5MOdV4uqZdtJpvZLRtUb8LQ=
Message-ID: <34d269d40809120932h3525bb97nbd8d982ff1aff501@mail.gmail.com>
Date: Fri, 12 Sep 2008 10:32:57 -0600
From: "Alex Hunsaker" <badalex@gmail.com>
To: "Zdenek Kotala" <Zdenek.Kotala@sun.com>
Subject: Re: hash index improving v3
Cc: "Tom Lane" <tgl@sss.pgh.pa.us>, "Xiao Meng" <mx.cogito@gmail.com>,
	pgsql-patches@postgresql.org, "Jonah H. Harris" <jonah.harris@gmail.com>
In-Reply-To: <48CA3312.1040508@sun.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
References: <10995.1220577198@sss.pgh.pa.us> <22574.1220728178@sss.pgh.pa.us>
	<34d269d40809061923k4c731481gf580d46d259defb0@mail.gmail.com>
	<20080909134839.GC6714@it.is.rice.edu>
	<34d269d40809091823t204e252fl246443be71aed9ff@mail.gmail.com>
	<34d269d40809092004q10b5da87r6fa3fe86926476ff@mail.gmail.com>
	<48C7DDF9.6020507@sun.com>
	<34d269d40809100927n20864e41p65c068c4ec6f6d74@mail.gmail.com>
	<34d269d40809102045t2f8037a4jfe862264f6f04454@mail.gmail.com>
	<48CA3312.1040508@sun.com>

On Fri, Sep 12, 2008 at 3:14 AM, Zdenek Kotala <Zdenek.Kotala@sun.com> wrote:
> Alex Hunsaker napsal(a):
>>
>> On Wed, Sep 10, 2008 at 10:27 AM, Alex Hunsaker <badalex@gmail.com> wrote:
>>>
>>> On Wed, Sep 10, 2008 at 8:47 AM, Zdenek Kotala <Zdenek.Kotala@sun.com>
>>> wrote:
>>>>
>>>> What locale did you use? It would be nice to have also comparing between
>>>> C
>>>> and any UTF8 locale. I think we should see big benefit when non C locale
>>>> is
>>>> used.
>>>
>>> Err yes this was UTF8, Ill see about doing a C locale.
>>>
>>
>> And here it with a C locale:
>>
>> pgbench -c1 -n -t100000 -f bench.sql
>> cvs head: tps = 5142.784262
>> v5:           tps = 6169.405965
>
>
> If I look on both results
>
>                C       UTF8    difference
> -----------------------------------
> cvs head:       5140    5050    -2%
> v5:             6170    5750    -7%
> improvement:    20%     14%
>
> than I little bit wonder. I personally expected bigger difference of UTF8
> comparing between CVS a v5. This test also shows that locale selection has
> bigger impact on performance in v5 case, but result is still better than cvs
> head.

Right, I think part of it is I need to try again with a larger
dataset... The reason I did three runs before was because it was
variable between (v5 between 5700 and 6200).  Also the query im doing
ming be two simplistic because it runs to fast, thats part of why I
get the variable speed between runs (i think...)

Suggestions?