maria-developers team mailing list archive

[Michael Widenius <monty@xxxxxxxxxxxx>]

Hi!

>>>>> "MARK" == MARK CALLAGHAN <mdcallag@xxxxxxxxx> writes:

<cut>

MARK> 703     1425    2593    4379    4176    4110    maria.nostat
MARK> 684     1281    2439    4064    3896    3882    maria.stat

MARK> oltp-ro run again for maria.nostat
MARK>                  761     1278    2593    4406    4207    4164    maria.nostat

Any idea why is this so different from the previous nostat ?

<cut>

MARK> 13418   24980   49777   94425   88083   86854   maria.nostat
MARK> 13062   23378   46636   83449   77714   77245   maria.stat

MARK> simple-allcols re-run for maria.nostat
MARK>                 12024   23891   50433   93778   88210   86127   maria.nostat

Same here.

MARK> So it looks like the regression at --num_threads==1 is real.

<cut>

Hakan, when you run tests, please make a test between MariaDB 5.1 and
MariaDB 5.2 and check if there is any performance differences when 5.2
is run without userstats enabled (default).

The only thing that I can think of that could cause any regression is
that wrapper around ha_read calls.  As these are small online functions I
can't see how the overhead could be notable.

Regards,
Monty