Re: setiathome on a UP1000

["Chris Eleveld (Sector 7 Group)" <Chris_Eleveld sector7 com>]

Well just from observed performance the Tru64 binary runs faster on my
UP2000 on the average.  It's a 750MHz EV67 chip.  It looks to average around
6 hours per work unit.  Though judging from strace I don't expect
significant performance with all the page faults and asking for the current
time.  Also I suspect there is a lot of memory I/O going on

----- Original Message -----
From: "Donsbach, Jeff" <Jeff.Donsbach@compaq.com>
To: <axp-list@redhat.com>
Sent: Sunday, May 05, 2002 8:24 AM
Subject: RE: setiathome on a UP1000


>
> Yes, really. Those "egcs" strings probably come from "glibc" functions
which are linked statically in the binary. The main code of the client
itself was built with "cxx".
>
> Actually, there could be a small performance loss on EV6 systems. The
"libcpml" linked into the binary was the EV5 version so as to allow it to
run on anything. It shouldn't be too significant though.
>
> Jeff D
>
> -----Original Message-----
> From: Stefan Schroepfer [mailto:whoknows@onlinehome.de]
> Sent: Friday, May 03, 2002 7:43 PM
> To: axp-list@redhat.com
> Subject: Re: setiathome on a UP1000
>
>  > There shouldn't be any substantial performance difference between the
>  > Linux client and the Tru64 client. They were both built with "cxx".
>
> Really? "strings" shows several occurences of "cxx" in the
> alpha-compaq-T64Uv4.0d binary, while the Linux client has none.
> In contrary, the linux client has 230 (!) occurences of the string
> "GCC: (GNU) egcs-2.91.66", while the Tru64 client has none.
>
> Regards,
> Stefan Schroepfer
>
>
>
> _______________________________________________
> Axp-list mailing list
> Axp-list@redhat.com
> https://listman.redhat.com/mailman/listinfo/axp-list
>
>
>
> _______________________________________________
> Axp-list mailing list
> Axp-list@redhat.com
> https://listman.redhat.com/mailman/listinfo/axp-list