| From: | "Jim C(dot) Nasby" <jnasby(at)pervasive(dot)com> |
|---|---|
| To: | Greg Stark <gsstark(at)mit(dot)edu> |
| Cc: | Luke Lonergan <llonergan(at)greenplum(dot)com>, Dann Corbit <DCorbit(at)connx(dot)com>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Simon Riggs <simon(at)2ndquadrant(dot)com>, pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: Merge algorithms for large numbers of "tapes" |
| Date: | 2006-03-09 16:35:52 |
| Message-ID: | 20060309163552.GD45250@pervasive.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Wed, Mar 08, 2006 at 10:20:08PM -0500, Greg Stark wrote:
> > For that to be of any use, wouldn't you need to use only as many tapes
> > as spindles/2? Otherwise you're still trying to read and write from the
> > same set of drives, which means you're probably doing a lot of seeking.
> > Or do the tape algorithms re-write data as they read it?
>
> Well, spindles-1. I was thinking as many tapes as you have spindles *in total*,
> ie, including the output tape. You only have one output tape for each n-way
> merge though.
Well, the reality remains though; most folks are unlikely to setup
enough dedicated temp areas so that we can do one tape per disk, so it
would be really good to have a sort method that didn't rely on that.
--
Jim C. Nasby, Sr. Engineering Consultant jnasby(at)pervasive(dot)com
Pervasive Software http://pervasive.com work: 512-231-6117
vcard: http://jim.nasby.net/pervasive.vcf cell: 512-569-9461
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Jonah H. Harris | 2006-03-09 16:35:55 | Proposal for SYNONYMS |
| Previous Message | Tom Lane | 2006-03-09 15:41:10 | Re: [COMMITTERS] pgsql: Remove Christof Petig copyright on include file, |