On 954517020 seconds since the Beginning of the UNIX epoch
"John A. Maier" wrote:
>
>> The '*Local* Communication bandwidths' numbers don't looks
>> rigth compared
>> to Simon's results for a ds20/OSF1. I can't believe NetBSD is
>> that bad :)
>
>Correct me if I'm wrong, but I seem to remember someone once posting that   
>code generated by DECs compiler on OSF/1 better/more efficient that GNUs.   
> I remember years ago, DEC talking about the Alpha, and how their   
>compiler takes advantage of specific Alpha optimizations to help code   
>execute more efficiently.

That isn't enough to explain the descrepancy on the DS20.  My
PWS433a has better local communications speeds than what Manuel
posted.  There must be some sort of problem.  Also note that Manuel
was getting abysmal mem bandwidth.  The compiler optimisation
shouldn't affect that too much.

 == Roland Dowdeswell                      http://www.Imrryr.ORG/~elric/  ==
 == The Unofficial NetBSD Web Pages        http://www.Imrryr.ORG/NetBSD/  ==
 == The NetBSD Project                            http://www.NetBSD.ORG/  ==