Dienstag, April 18, 2006

Axapta Performance Comparison Oracle 9i and SQL2005

I finished the first performance tests and here are the results. The configuration is as follows:
Oracle:, used Axapta client is 1951.3733
SQL2005: used Axapta client is Kernel Rollup 1

The test server is a P4 2.8 GhZ with 2GB of RAM. Both Oracle and SQL2005 have 1GB RAM assigned (not at the same time. Either Oracle OR SQL2005 is running.)

The first tests show that SQL2005 is about 25-30% slower than Oracle if the database is running in SQL2000 mode.
The performance is a little bit better when running in native SQL2005 mode: then SQL2005 is about 14-19% slower than Oracle.

The biggest surprise was that Oracle was always faster when a certain test (form filter/report execution) was done the second time and the data was coming from cache. It seems that the caching algorithm of Oracle was much better than from SQL2005. There was a range of 45-150% worse performance in SQL2005 when getting data from cache.

Keine Kommentare: