Space Forum / SETI / May 2004

Graham Vincent adjusted his tin foil beanie and asbestos underwear to
write:

From the xsetiathome readme:-

<------------snip------------>

KNOWN BUGS
       Please note, this xsetiathome binary is an initial experiment.
       We are fully aware that it may have some significant bugs and
       we are continuing to work on it.  Please do not report bugs
       on this binary at this time.

       Only one instance of 'setiathome -graphics' can be operating at
       any one time on a system.  There is no conflict resolution with
       the allocated shared memory segments and semaphores.  Therefore,
       only one instance of xsetiathome is practical.

       Outstanding shared memory segments and semaphores may be left
       active in case of an abnormal exit of the 'setiathome -graphics'
       process.  These can prevent any future invocation of
       'setiathome -graphics'.  To resolve this problem, use 'ipcs'
       and 'ipcrm' to remove shared memory segments and semaphores that
       are not associated with a process.  This behavior may vary
       depending upon how your UNIX system handles this situation.
       See also ipcs(1) and ipcrm(1)

       The xsetiathome process may affect system performance adversely.
       This behavior may vary depending upon your version of the
       X11 Windowing system client libraries and/or your X server
       and its operating mode/visual depth.

<------------pins------------->

HTH

I noticed that just running the X interface (tk-pppoe) for rp-pppoe used
something like 20% cpu time on a K6-2/400 (kernel pppoe used no cpu time).  
So if just monitoring or graphing network speed takes that much cpu, I
imagine something more graphic intensive, like xsetiathome would use more.  
But I run SET@home -nice 19 headless on K6-2/400 PC and Celeron 300
pppoe/firewall/smtp/webserver, where it runs unobtrusively (not even
noticed) without unnecessary graphics.