Nice value worthless??? / Renice?

[David F. Carlson]

In article <1800006 at spdyne>, root at spdyne.UUCP writes:
> I have a question about the nice values of processes:
> 	Now I'm running a Compaq 386/20, with (sorta slow drives (40ms)),
> and only 3 Meg of ram.  Microport UNIX with DOS-MERGE. (Sys V)
> 
> As far as I can tell, the nice value does absolutly nothing!
> 
> 	Is it just the way that Sys V works or is it just that I don't have
> squat for ram?  (Merge wants 2.6 Meg), Slow disks?  (I know that the
> CPU waits on the transfer - right?)

Nice affects the scheduler priority.  It does not deal very well with the
lower level system performance issues like "do I have more than 1 process
able to run", etc.  This is important with a machine with small RAM and
a slow disk.  Swapping or page faulting a running process blows it out
for approx. that 40ms of drive seek time.  No RAM with a large make (read
at least 6 processes that must run to do your compile) can make for swap
or DPVM thrash VERY easily.  If your vi is paged out, it will not run until
the page it needs make it through the diskIO queue, which can seem like
forever.

In summation, if a job is IO or memory (or both) bound, don't expect nice,
(which is really affects CPU advantage) to do squat for you.

> 	Also, Does anyone have a 'renice' for Sys V?  I have one for BSD, but
> of course it is worthless under Sys V. 
> 	-Chert Pellett
> 	 root at spdyne

One of the UNIX source groups has renice in its archives.  Look there.


-- 
David F. Carlson, Micropen, Inc.
micropen!dave at ee.rochester.edu

"The faster I go, the behinder I get." --Lewis Carroll