Fun with - today: the SAP SD benchmark

Why did i wrote the article about the “non-unicode/unicode” benchmarketing? Because of articles like this one: SAP-Benchmarkergebnisse für Intels neuen Hexa-Core-Prozessor. It´s frightning. The publications of the publisher Heise had a good reputation in the past. But they jumped the shark long time ago. They should really stay at testing displays and notebooks and something like that and Andreas Stiller should transform his “Prozessorgefluester” column into an independent blog. One of the highlights of the last few day - “Don´t understand the nature of benchmarks”:

Bei den Vierprozessorsystemen konnte sich HPs ProLiant DL580 G5 mit Windows Server 2003 und Microsoft SQL Server 2005 mit 5155 SD-Usern(1,97 s, 25.380 SAPS) klar vor Suns Fire X4450 mit Solaris 10 und MaxDB 7.6 behaupten, das nur auf 4600 SD-User (1,94s, 23120 SAPS) kommt.

As i wrote before, the HP result of 25380 SAPS is a non-unicode result, the Sun result is based on a unicode installation. So dear Before drawing any conclusions from existing data, try to understand the data.