trouble with "nice" increments

arritt at kuhub.cc.ukans.edu arritt at kuhub.cc.ukans.edu
Wed Jul 4 02:36:19 AEST 1990


I am using a 4D/25G running IRIX 3.2.  The "nice" command does not seem
to do what I would like; i.e., establish varying priorities for different
jobs.  Specifically it does not seem to respond to the priority increment
argument.  As a test, I started three identical jobs, as follows:

JOB 1:  ascot.exe < ascot.dat > ascot.out &
JOB 2:  nice -19 ascot.exe < ascot.dat > ascot.nice &
JOB 3:  nice ascot.exe < ascot.dat > ascot.nice2 &

I then checked the jobs every so often, and found the following cpu usage:

JOB 1 (change from previous)        JOB 2 (change)         JOB 3 (change)
4:34                                1:49                   0:01
4:51  (0:17)                        2:07  (0:18)           0:16  (0:15)
5:13  (0:22)                        2:30  (0:23)           0:35  (0:19)
5:41  (0:28)                        2:57  (0:27)           0:57  (0:18)

As you can see from the (change) entries, JOB 1 and JOB 2 used the same cpu
within roundoff error to one second, while JOB 3 uses less cpu.  I would have
expected for JOB 2 to get the least cpu (at "nice -19"), JOB 3 to get a little
more (since "nice" should default to "nice -10") and JOB 1 to get the most.

Each job is 397 blocks (about 1.6 MB) so all three jobs should fit easily into
the system's 16 MB of physical memory.

What am I doing wrong?  Any help you can give would be most appreciated.
________________________________________________________________________
Ray Arritt                        | 
Dept. of Physics and Astronomy    |
Univ. of Kansas                   |
Lawrence, KS  66045               |
arritt at kuhub.cc.ukans.edu         |
arritt at ukanvax.bitnet             |
                               



More information about the Comp.sys.sgi mailing list