On Mon, 26 Jan 2009, Shane R. Spencer wrote:
> A context switch is faster than running ssh and checking for exit status
> and then immediately running nfs-kernel-server right after it. and so on.
I think you're missing the point. Not running these new fancy init systems
I don't know how they're verifying dependent services (wouldn't it be much
simpler to to spawn a subshell and to a 'status' call on the requisite init
scripts?), but that's not germane to multitasking 101. Multitasking on a
single core (assuming all other things being equal) will never be faster
than multitasking on multiple cores -- unless your particular task is
extremely sensitive to cache latencies. But I hardly think that applies
here.
The only circumstance that I could see problems parallelizing service
launches would be if several of those launches were extremely disk I/O
intensive. Barring that, it's a safe bet that users on a multicore system
will see significantly better performance than those on a single core
system.
--Arthur Corliss
Live Free or Die
---------
To unsubscribe, send email to <aklug-request@aklug.org>
with 'unsubscribe' in the message body.
Received on Mon Jan 26 10:59:12 2009
This archive was generated by hypermail 2.1.8 : Mon Jan 26 2009 - 10:59:12 AKST