Wednesday, June 6, 2007

Benchmarking Idiocy

So, a "person who shall not be named in my blog" posted a link to this story over at DailyTech. Now, a few notes on the article (for people who can't read):
a.) There is a decided clock speed DISADVANTAGE between the Barcelona and the Xeon 3220 (1.6ghz vs. 2.4ghz)
b.) The Intel Xeon 3220 is running on the newest Intel Chipset (x38) versus an older, non-HT3 compliant nVidia Prof. 3400
c.) The Intel Xeon 3220 is running 4gb of DDR2-800 versus 4gb of DDR2-667 on the AMD Barcelona.
d.) The Intel Xeon 3220 is a "mature" processor built on already established silicon that's been out for at least a year. The Barcelona is pre-release silicon that hasn't even completed the final silicon runs at the foundry.

So, people who claim that Barcelona isn't performing well are really basing their ridiculous assumptions on pre-release, heavily biased benchmarking.

Note: Kristopher Kubicki and I do have a good history of interaction together and I know that this article is not to belittle AMD's work so far. The points I've enumerated above simply point to 4 KEY discrepancies in the benchmarking methodology that people should be aware of before jumping to conclusions.

Technorati Tags: , , , , ,

Powered by ScribeFire.

No comments: