@ Downix
I can ask around to find a 3rd party willing to make one, if you'd back me up on it being a fair and unbiased test.
Be my guest. These tests should be well documented and easily reproducable and I would happily accept them as being genuine.
You're the one challenging it.
So you think I should blindly accept everything BB feeds us? As I said, based on my personal experiences, the performance of currently available G3 solutions are about equivalent.
With regard to processor performance tests, I rather believe the benchmarks provided by IBM and Motorola. Do you truly understand what it would take to objectively test all possible benchmark scenarios? For instance did these "tests", BB refers to, include Altivec optimised software tests? Altivec optimisations can result into performance gains up to 60%, but the performance gains do differ for different types of software.