On Mon, Sep 23, 2013 at 12:47 AM, Alexander Korotkov
<aekorot...@gmail.com>wrote:

> It's probably an option to select 64 entries instead of 32.
> There is still some regression in update speed. However, there is also
> room for improvement patch. It searches item index entries 2 times on
> insert: in dataLocateLeafItem and dataPlaceToPage. We can save full results
> of dataLocateLeafItem, but it require some rework of gin btree interface:
> store not only offset of item.
>

In the attached version of patch double finding of ItemPointer during
insert is avoided. Overhead becomes lower as expected.

         event         |    master       |  16-entries     |  32-entries
  |  64-entries     |  128-entries    |
-----------------------+-----------------+-----------------+-----------------+-----------------+-----------------+
 index_build           | 00:01:50.042658 | 00:01:54.130873 | 00:01:59.37302
 | 00:01:55.959693 | 00:01:58.126407 |
 index_build_recovery  | 00:00:19        | 00:00:06        | 00:00:06
 | 00:00:06        | 00:00:06        |
 index_update          | 00:05:18.215707 | 00:05:38.40231  |
00:05:30.658786 | 00:05:27.664312 | 00:05:30.815876 |
 index_update_recovery | 00:01:48        | 00:01:53        | 00:01:50
 | 00:01:44        | 00:01:46        |
 search_new            | 00:25:21.481699 | 00:23:20.324152 |
00:24:02.120438 | 00:22:50.989723 | 00:23:05.703824 |
 search_updated        | 00:25:57.622592 | 00:26:43.531979 |
00:26:08.003085 | 00:24:36.669028 | 00:26:09.175243 |


     label     |    size    | 16-entries | 32-entries | 64-entries |
128-entries |
---------------+------------+------------+------------+------------+-------------+
 new           |  884514816 |  417013760 |  421240832 |  430350336 |
450994176 |
 after_updates | 1595252736 |  711368704 |  719380480 |  735682560 |
774275072 |

------
With best regards,
Alexander Korotkov.

Attachment: gin-packed-postinglists-6.patch.gz
Description: GNU Zip compressed data

-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Reply via email to