Hi!
On Mar 16, 2009, at 12:24 PM, MARK CALLAGHAN wrote:
Can you give more details on the case where pool-of-threads makes
anything faster? I have no problem making sysbench throughput 2X
The only case where it works well is when you have lots of primary key
lookups with no transactions. Otherwise? Its junk.
Eric is working on a new implementation bit once he finishes with the
new protocol work (if you are interested in the new protocol work... I
believe Monty/etc are pulling it into MariaDB as well, so you might be
able to patch off of them).
BTW the multi-threaded scheduler in Drizzle benchmarks better then the
MySQL multi-threaded. You might be able to pick out our scheduler
interface and pull it into MySQL if you start with a 5.1 tree. The
changes required are not all that large to make it loaded (MontyW had
done some work on this before he gave up on the main MySQL tree).
Cheers,
-Brian
--
_______________________________________________________
Brian "Krow" Aker, brian at tangent.org
Seattle, Washington
http://krow.net/ <-- Me
http://tangent.org/ <-- Software
_______________________________________________________
You can't grep a dead tree.
_______________________________________________
Mailing list: https://launchpad.net/~drizzle-discuss
Post to : [email protected]
Unsubscribe : https://launchpad.net/~drizzle-discuss
More help : https://help.launchpad.net/ListHelp
