Red Hat Enterprise Linux Server release 6.5 (Santiago)

Compare to mysql 5.1 this query lost 42s for about 300s execution time . Removing the GROUP BY, Sebastien found that the regression can not be reproduce.

cnf file

Not the same server MariaDB supposed to be faster :
Mysql :
CPU : 2 (2Ghz)
Mem : 4Go

MariaDB :
CPU : 4 (2,6Ghz)
Mem 16Go.

Voila déjà les résultat sur mariaDB10 en pièce jointe.

Pour avoir une machine ISO pour la version 51 de mysql, je dépend du système et ne peut te fournir les informations de suite.

Hello,
With Stephane we have compare the follow query :
create temporary table toto engine=blackhole select a13.id_vehicule id_vehicule,
a12.id_energie id_energie,
a11.num_jo_dans_mois WJXBFS1,
a11.num_jo_dans_annee WJXBFS2
from `T_CALENDRIER` a11
cross join `T_ENERGIE` a12
cross join `T_VEHICULE` a13
where (a11.Id_Calendrier in (20131031)
and a13.id_genre in (45));

On Mysql this query execute in 3510ms
on MariaDB10 the query takes 4633ms.

Then we try benchmarck :
on Mysql :
/[16:14:16][ 203 ms]/ select benchmark(1000000, encode('hello', 'goodbye') );
/[16:15:39][ 187 ms]/ select benchmark(1000000, 1024/8 );
On MariaDB :
/[16:13:56][ 453 ms]/ select benchmark(1000000, encode('hello', 'goodbye') );
/[16:15:53][ 359 ms]/ select benchmark(1000000, 1024/8 );

Its seems we have a problem with our server.

Thx for your help

Closing for now. Please comment to reopen if you have second thoughts regarding the cause of the problem.