Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 05 Oct 1996 03:45:22 -0500
From:      "Chris Csanady" <ccsanady@friley216.res.iastate.edu>
To:        Peter Wemm <peter@spinner.dialix.com>
Cc:        smp@freebsd.org
Subject:   Re: top patches for smp.. 
Message-ID:  <199610050845.DAA03603@friley216.res.iastate.edu>
In-Reply-To: Your message of Sat, 05 Oct 1996 13:25:32 %2B0800. <199610050525.NAA00432@spinner.DIALix.COM> 

next in thread | previous in thread | raw e-mail | index | archive | help

>If anybody's interested, these patches will show some extra info when
>top is running.  I can't remember if I've send this already..  Hope
>not. :-)
>
>With this, instead of "RUN", you will see either "CPU0/1" or "RUN/1".
>In this case, "CPU0/1" means that the process is currently on cpu 0,
>and the "1" is meaningless.. It's set at the same time I just realised.
>:-]  The idea was that you could see the current and "last" cpu, so you
>had a chance of seeing procs bouncing from one cpu to another, maybe
>later..
>
>"RUN/1" means it is runnable, waiting in the run queue and last ran on
>cpu#1.

Hmm.. when I run this, i the states are always either "RUN/0", or "CPU#0/".
(no 1 at the end of that.)  Am I not using both cpus?  I could swear I am
since a parallel make is about 70% faster.. :\  Although, i hardly ever
see a "CPU#0/".

Ive been trying to think why such a thing would occur, but so far have come
up with nothing.  Something I have been wondering about though.. shouldn't clock
interrupts be processed on both cpus?  I mean isnt this where the p_cpuest
is incremented for the current process?  It seems that if only one cpu can be
in kernel then only one would handle it, and it would mess with the scheduling
priorities.  But then, I may just have no clue what im talking about. :)  Im
not quite sure about how interrupts need to/are handled now..  Anyways,
its just so hard to think about some of this stuff..

Another thing i was pondering..  if the recomputation of priorities (among
other fixed time interval things) happened to coincide on both cpus,
undesireable effects may occur..

Well, i really need to get some sleep..

Chris Csanady




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199610050845.DAA03603>