David Kanter (dkanter@realworldtech.com) on 11/29/05 wrote:
---------------------------
>an (an@yahoo.com) on 11/29/05 wrote:
>---------------------------
>>I wonder what the more critical posters around here would >say about, lets say Sun,
>>flying reviewers from a few sites out to a site to review a >future product?
>
>No comment.
>
>>Especially
>>if all the reviews came out with striking similarities in >terminology (throughput
>>computing everywhere!) and benchmarks used (especially one >previously unheard of
>>benchmark which strangely enough make their new product >really shine, especially
>>if you get the feeling the reviewer(s) had no idea what it >was doing).
>
>Did you really get the idea that I didn't have a clue what any of the benchmarks
>were doing? I honestly thought I explained them relatively well.
>
>DK

Well first, I ws referring to Black-Scholes, and it was not necessarily directed solely at you. The fact that other sites used the same benchmark, and the descriptions were fairly cookie cutter made me wonder if the reviewers were not just given a fairly obscure benchmark with a description and ran with it. After all how familiar are you (they) with the input and output such a program normally uses or produces? Were you able to run it under different scenarios (other than just changing the # of iterations and # of threads)?