Re: Compression related performance

Neal Richter <[email protected]> Sat, 15 Jan 2005 16:06:09 -0700 (MST)
Newsgroups gmane.comp.web.htdig.devel
Message-ID <[email protected]>
Jim,
 	Thanks for the numbers.  Pretty different from Joe's.

Question #1)

    Could you download a fresh copy of BDB 3.0.55 from
http://www.sleepycat.com/download/index.shtml

It's easy to build, I am most interested in the results when you run 
db_stat on the uncompressed files and report back the fill-factor.  This 
is a measure of wasted space in the BDB pages.  This won't work on the 
compressed index.

Question #2)
   What page_size are you using?

Question #3)
   Do you ever see any errors when using the compressed index?

   Joe:  Do you?

Thanks

Neal

On Fri, 14 Jan 2005, Jim wrote:

> Here is another performance comparison collected when the server was under 
> less load. The data is essentially the same as that used for the last set of 
> numbers.
>
> Uncompressed
> ------------
> 416.28 user
> 97.72 system
> 1:30:23 elapsed
>
>  10510336 Jan 15 00:00 db.docdb
>  10444800 Jan 15 00:00 db.docs.index
>  90488832 Jan 15 00:00 db.excerpts
>  436043776 Jan 15 00:00 db.words.db
>
> Compressed
> ----------
> 2008.63 user
> 133.35 system
> 2:40:16 elapsed
>
>  10526720 Jan 14 09:35 db.docdb
>  10444800 Jan 14 09:35 db.docs.index
>  90611712 Jan 14 09:36 db.excerpts
>  113773568 Jan 14 09:36 db.words.db
>     16384 Jan 14 09:36 db.words.db_weakcmpr
>
>
> I do have a large data set that I could test with if it would be of real
> value. However I would need to try rebuilding ht://Dig with large file
> support before I could consider dropping compression for that set.
>
>
> Jim
>
>
> -------------------------------------------------------
> The SF.Net email is sponsored by: Beat the post-holiday blues
> Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
> It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
> _______________________________________________
> ht://Dig Developer mailing list:
> [email protected]
> List information (subscribe/unsubscribe, etc.)
> https://lists.sourceforge.net/lists/listinfo/htdig-dev
>

-- 
Neal Richter 
Knowledgebase Developer
RightNow Technologies, Inc.
Customer Service for Every Web Site
Office: 406-522-1485



-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
_______________________________________________
ht://Dig Developer mailing list:
[email protected]
List information (subscribe/unsubscribe, etc.)
https://lists.sourceforge.net/lists/listinfo/htdig-dev