On Thu, 9 Mar 2006, Christian Smith wrote:
>On Wed, 8 Mar 2006 [EMAIL PROTECTED] wrote:
>
>>
>>One question though: are the file access "sorted", so that seeks are
>>minimised when performing a transaction (making the assumption that the
>>file is not fragmented on disk)?
>
>
>The OS will sort IO into whatever order it sees fit. Reads will be done in
>order, and writes will be likely be written to the disk in ascending disk
>block order. SQLite can do little to affect this policy.
If you're on Linux, and use ext3 (or Reiser4, I believe) you can mount
filesystems that log data to the journal on sync using "data=journal"
mount option. This may benefit SQLite in writing all updated data blocks
to the contiguous journal before the final in-place update on the device.
The speed benefit can be substantial IIRC, as the fsync is considered
complete once the data reaches the journal.
>
>
>>
>>Nicolas
>>
>
>Christian
>
>
--
/"\
\ / ASCII RIBBON CAMPAIGN - AGAINST HTML MAIL
X - AGAINST MS ATTACHMENTS
/ \
