Re: on-disk bitmap indexes

Lists: pdxpug
From: Mark Wong <markw(at)osdl(dot)org>
To: pdxpug(at)postgresql(dot)org
Subject: on-disk bitmap indexes
Date: 2006-09-25 21:56:32
Message-ID: 45185090.3030203@osdl.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pdxpug

I remember hearing some interest about on-disk bitmap indexes. In case
no one has seen it, just wanted to forward a Greenplum presentation
about their implementation that's been submitted to PostgreSQL:

http://files.meetup.com/230033/BizgresOnDiskBitmapIndex.pdf

Mark


From: "David E(dot) Wheeler" <david(at)kineticode(dot)com>
To: Mark Wong <markw(at)osdl(dot)org>
Cc: pdxpug(at)postgresql(dot)org
Subject: Re: on-disk bitmap indexes
Date: 2006-09-26 03:41:30
Message-ID: A93F55F1-FBD0-4CA4-806D-CD611DF4D134@kineticode.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pdxpug

On Sep 25, 2006, at 14:56, Mark Wong wrote:

> I remember hearing some interest about on-disk bitmap indexes. In
> case no one has seen it, just wanted to forward a Greenplum
> presentation about their implementation that's been submitted to
> PostgreSQL:
>
> http://files.meetup.com/230033/BizgresOnDiskBitmapIndex.pdf

I don't really understand it, but their benchmark graphics sure make
it look hella impressive!

Best,

David


From: Mark Wong <markw(at)osdl(dot)org>
To: "David E(dot) Wheeler" <david(at)kineticode(dot)com>
Cc: pdxpug(at)postgresql(dot)org
Subject: Re: on-disk bitmap indexes
Date: 2006-09-26 17:11:44
Message-ID: 45195F50.5040501@osdl.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pdxpug

David E. Wheeler wrote:
> On Sep 25, 2006, at 14:56, Mark Wong wrote:
>
>> I remember hearing some interest about on-disk bitmap indexes. In
>> case no one has seen it, just wanted to forward a Greenplum
>> presentation about their implementation that's been submitted to
>> PostgreSQL:
>>
>> http://files.meetup.com/230033/BizgresOnDiskBitmapIndex.pdf
>
> I don't really understand it, but their benchmark graphics sure make it
> look hella impressive!

Wouldn't it be fun to go through a tuning exercise using the TPC-H? :)

Mark


From: "David E(dot) Wheeler" <david(at)kineticode(dot)com>
To: Mark Wong <markw(at)osdl(dot)org>
Cc: pdxpug(at)postgresql(dot)org
Subject: Re: on-disk bitmap indexes
Date: 2006-09-26 17:33:42
Message-ID: 972C6F0E-0925-47E1-8E4D-13D28AC5A367@kineticode.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pdxpug

On Sep 26, 2006, at 10:11, Mark Wong wrote:

> Wouldn't it be fun to go through a tuning exercise using the TPC-H? :)

The what?

Best,

David


From: Mark Wong <markw(at)osdl(dot)org>
To: "David E(dot) Wheeler" <david(at)kineticode(dot)com>
Cc: pdxpug(at)postgresql(dot)org
Subject: Re: on-disk bitmap indexes
Date: 2006-09-26 17:47:32
Message-ID: 451967B4.5020002@osdl.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pdxpug

David E. Wheeler wrote:
> On Sep 26, 2006, at 10:11, Mark Wong wrote:
>
>> Wouldn't it be fun to go through a tuning exercise using the TPC-H? :)
>
> The what?

The TPC-H is an ad hoc decision support benchmark.
http://www.tpc.org/tpch/default.asp

I just noticed the presentation referenced dbt3, not the TPC-H but dbt3
is an implementation of the TPC-H. The whitepaper referenced at the end
of the presentation (login required to d/l) references TPC-H instead of
dbt3.

Mark


From: Selena Deckelmann <selena(at)chrisking(dot)com>
To: pdxpug(at)postgresql(dot)org
Subject: Re: on-disk bitmap indexes
Date: 2006-10-10 17:59:13
Message-ID: 906f8820aefa5cfa8bb36b7c1e6f8822@chrisking.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pdxpug

On Sep 25, 2006, at 2:56 PM, Mark Wong wrote:

> I remember hearing some interest about on-disk bitmap indexes. In
> case no one has seen it, just wanted to forward a Greenplum
> presentation about their implementation that's been submitted to
> PostgreSQL:
>
> http://files.meetup.com/230033/BizgresOnDiskBitmapIndex.pdf

I finally read this today. The performance graphs are impressive. I
would like someone to explain slide #8 to me. I get the general idea
of binning data but I need some hand-holding on how the
equality/interval/range encoding works. The arrows are not doing much
for me.

-selena