[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
R Clint Whaley wrote:
> >Very well then, can someone in the ATLAS community kindly do a
> >nonpartisan comparison?
> You are the one publishing numbers. Either correct them, or remove them.
> All it should take is a relink and a rerun.
so we went ahead and reran all the experiments with the
PIII 256K kernel that you sent. No noticeable difference.
Indeed, we now seem to do better compared to ATLAS...
Given that some have been freely accusing people of performance
irregularities, we did to a hand-recount of all machine cycles.
Since this was only done in a county which has close ties to
one of the parties, I strongly suggest that you check with the
secretary of state and insist that this recount was not
performed by the set deadline. This will allow us to then
dispatch our lawyers and all will probably be resolved by
SC2001. We would like to offer a compromise: It would
be good if someone would do a hand-recount in Knox County
as well, while we wait for the overseas verdict.
New performance numbers are on the webpage.