"To answer this cacophony of angry readers, Hardware Central followed the initial article several weeks later with another piece - this time backed by benchmarks. Very sadly, these tests were excruciatingly slanted toward RDRAM, while handicapping SDRAM. The Rambus endowed system had the most expensive and rare variety of RDRAM, PC800, which is particularly onerous since most RDRAM systems come populated with slower and cheaper PC700 RDRAM. At the same time, the Via system was burdened with slow CL3 SDRAM, while the true contender, the Intel 440BX based system, was sabotaged by forcing it to run its SDRAM and processor bus at a much slower 100 MHz.
Even despite these machinations, the test results show effectively a performance wash-out, despite the extreme price premium for the RDRAM system. Somehow, in spite of the sagging (and, in actuality, invalid) results, the conclusion is resoundingly upbeat and attempts to paint their test findings as conclusive confirmation of RDRAM's great promise.
Perhaps as a testimonial to the character of Rambus Inc. itself, this dubious article appears several times as a link on Rambus's corporate site as a presumably relaible independent review of their technology.
Even our friend Anand Lal Shimpi of AnandTech wrote an uncharacteristically soapy eyed article on this technology recently that seemed to closely follow the Rambus company line. To his credit, however, Anand has followed this questionable article with a new one backed with benchmarks which largely substantiate the condemning results that Dr. Thomas Pabst derived months ago."