On Thu, Jan 13, 2022 at 06:17:54PM -0800, Jonathan Thornburg wrote:

> I've just noticed something odd about the scheduling of processes with
> varying 'nice' values (7.0-stable/amd64, GENERIC.MP): it appears that
> processes with 'nice 20' are given more favorable scheduling than those
> with 'nice 10', which is exactly the opposite of what I'd expect based
> on the man page for setpriority(2), "lower priorities cause more favorable
> scheduling" (and longstanding Unix experience).
> 
> In more detail:
> Right now I have 5 CPU-bound processes running (all the same binary, but
> with different command-line arguments and started from different working
> directories), on hardware with 4 CPUs visible to OpenBSD (quad-core Intel
> i7-8650U processor; hyperthreading is disabled both in the BIOS and by
> default in OpenBSD).  Of those 5 processes, 3 are at 'nice 20', and the
> other 2 are at 'nice 10'.  I expected the 2 'nice 10' processes to each
> get more CPU time than the 3 'nice 20' processes, but 'top -S -i -s1'
> shows exactly the opposite behavior: the 3 'nice 20' processes are each
> getting MORE CPU time (about 100% of a CPU each) than the 2 'nice 1
> ' processes (about 50% of a CPU each):
> 
> load averages:  5.04,  4.99,  4.04               gold.bkis-orchard.net 
> 18:02:38
> 176 processes: 4 running, 168 idle, 4 on processor             up 7 days, 
> 17:40
> CPU0:  0.0% user, 97.0% nice,  0.0% sys,  0.0% spin,  3.0% intr,  0.0% idle
> CPU1:  1.0% user, 99.0% nice,  0.0% sys,  0.0% spin,  0.0% intr,  0.0% idle
> CPU2:  0.0% user,  100% nice,  0.0% sys,  0.0% spin,  0.0% intr,  0.0% idle
> CPU3:  1.0% user, 99.0% nice,  0.0% sys,  0.0% spin,  0.0% intr,  0.0% idle
> Memory: Real: 5686M/13G act/tot Free: 1929M Cache: 6785M Swap: 0K/34G
> 
>   PID USERNAME PRI NICE  SIZE   RES STATE     WAIT      TIME    CPU COMMAND
> 95910 jonathan 104   20   39M   42M onproc/3  -        17:38 99.02% smp-O3
> 58006 jonathan 104   20   39M   42M run/2     -        42:45 98.97% smp-O3
> 63085 jonathan 104   20   39M   42M run/1     -        12:39 97.66% smp-O3
> 36985 jonathan  84   10   21M   24M onproc/0  -         5:57 49.66% smp-O3
> 95125 jonathan  84   10   21M   24M run/0     -        11:53 49.32% smp-O3
> 64031 _firefox  28    0  959M 1057M run/2     -        63:28  0.83% 
> firefox-esr
> 77428 _firefox   2    0 1381M 1355M sleep/2   poll    179:14  0.10% 
> firefox-esr
> 
> Am I missing something obvious?
> 
> --
> -- "Jonathan Thornburg [remove color- to reply]" <jthorn4...@pink-gmail.com>
>    on the west coast of Canada, eh?
>    "There was of course no way of knowing whether you were being watched
>     at any given moment.  How often, or on what system, the Thought Police
>     plugged in on any individual wire was guesswork.  It was even conceivable
>     that they watched everybody all the time."  -- George Orwell, "1984"
> 

Are youre processes multithreaded?? Check with top -H.

        -Otto

Reply via email to