Jairo - 2012-09-10

Hi all,

I tried to improve my Wikipedia Miner instance but the Waikato University demo is twice as fast as mine.

I tried following configuration:

- Amazon m2.2xlarge instance:
34.2 GB of memory
13 EC2 Compute Units (4 virtual cores with 3.25 EC2 Compute Units each)
850 GB of instance storage
64-bit platform
I/O Performance: High

- Tomcat 7 with Java 7:
JAVA_OPTS = "-Xms8g -Xmx34g -XX:+UseAdaptiveSizePolicy -XX:+AggressiveHeap"
CATALINA_OPTS = "-Xms8g -Xmx34g -XX:+UseAdaptiveSizePolicy -XX:+AggressiveHeap"

- Wikipedia Miner:
In configuration file named wikipedia.xml:

        <databaseToCache priority="speed">pageLinksIn</databaseToCache>
        <databaseToCache priority="speed">label</databaseToCache>

To cache frequently used files in the database

And nothing else.

I do not know what else I can do to improve performance. I need better performance. This is critical in my project.

Any ideas? Does anyone know a better setup?

Help me, please!

Thanks you in advance!,