db-ojb-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Armin Waibel <ar...@code-au-lait.de>
Subject [OTM] Massive performance decrease between Feb and Apr version
Date Sat, 10 Apr 2004 11:40:22 GMT
Hi all,

I made my regular performance measurement and stumple across a massive 
performance decrease of the OTM-api:

"ant perf-test" results against in-memory hsql

          |10-Apr |05-Feb
------------------------
jdbc     | 100%  | 100%
PB-api   | 124%  | 132%
ODMG-api | 230%  | 235%
OTM-api  | 329%  | 193%

You can see performance of PB- and odmg-api is constant.
20-30% overhead in the PB-api compared with native JDBC is a very good 
result. ODMG-api is 2,3 times slower than native JDBC that's not good 
but pessemistic locking is expensive and we never do a real refactoring 
on this implementation.

But the perfromance of the OTM-api compared with native JDBC decrease 
from 193% to 329%!! Means OTM is 3,2 times slower than native JDBC (from 
great 1,9 in Feb).
I think this is unacceptably, because if we use JDO or ODMG on top of 
OTM the performance will decrease a little more, maybe to 3,5-4,2 times 
slower than native jdbc.

I know "only blockheads look for performance only", but I think the 
overhead of OTM is a tad big ;-)

regards,
Armin


---------------------------------------------------------------------
To unsubscribe, e-mail: ojb-dev-unsubscribe@db.apache.org
For additional commands, e-mail: ojb-dev-help@db.apache.org


Mime
View raw message