On 08/06/2011 06:50, Alessandra Forti wrote:
> Hi Chris,
>> You seem to be suggesting we are publishing incorrectly. I believe
>> that what we are publishing is correct and self consistent. Since you
>> believe otherwise, can you explain what you think we are doing wrong.
>
> one of the reasons CPUAtlas was introduced and deemed better was
> because it solved QMUL discrepancies with atlas when you had in apel
> twice the number of raw cpu hours and published haf the HS06. There
> was a factor 2 that could be moved around. It seems you have corrected
> that now.
>
> cheers
> alessandra
|