|
OK, here is an explanation.
When dynamic priority scheduler is turned on, job's priority is no longer a
constant.
Well, external priority is a constant, but internal dispatching priority is
not.
When job uses CPU, its internal priority is gradually reduced (numerically
increased) - to increase the chance for other jobs to use CPU. When job
misses its chance to use CPU, because of higher priority jobs, its priority
is gradually increased (numerically reduced) - to avoid CPU starvation.
You may see this effect looking at WRKSYSACT screen - long running jobs
will show higher priorities - say 35 instead of 20 for typical interactive
jobs.
At timeslice end, job's priority is reset to initial value.
So in a sense, lower timeslice jobs are indeed more "reactive", because
their priority is reset more often.
To see this effect, you actually need to run non-interactive type of work
in interactive job.
For real "interactive" interactive job, the rules will be more complicated.
This is a much simplified explanation. There was a redbook SG24-4735-00
"AS/400 Performance Management V3R6/V3R7"
published in March 1997, which has a chapter on this with more detailed
explanation.
The purpose of all these complexities is to make CPU allocation more "fair"
and to prevent high-priority jobs from hogging CPU.
Alexei Pytel
"The better is the worst enemy of the good"
"Leif Svalgaard"
<leif@leif.org> To: <midrange-l@midrange.com>
Sent by: cc:
midrange-l-admin@mi Subject: Re: Timeslices
drange.com
08/29/2001 11:16 PM
Please respond to
midrange-l
turning off the dynamic scheduler, makes everything much more even.
Now there is no difference. All jobs proceed at the same rate.
----- Original Message -----
From: Alexei Pytel <pytel@us.ibm.com>
To: <midrange-l@midrange.com>
Sent: Wednesday, August 29, 2001 10:25 PM
Subject: Re: Timeslices
>
> It might happen because of dynamic priority scheduler.
> Before going into lengthy explanations of my theory which might be wrong
-
> could you set system value QDYNPTYSCD to 0 (off) and repeat you run.
> If effect you see will disappear, I might be able to explain it.
>
> Alexei Pytel
> "Leif Svalgaard"
>
> I have noticed that (interactive) jobs with a small timeslice
> are more "reactive" than jobs with a large timeslice.
> This seems to indicate that the OS/400 is not "truly"
> preemptive. Is this observation correct, or am I missing
> something, or should I even know?
_______________________________________________
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@midrange.com
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/cgi-bin/listinfo/midrange-l
or email: MIDRANGE-L-request@midrange.com
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.
As an Amazon Associate we earn from qualifying purchases.
This mailing list archive is Copyright 1997-2025 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].
Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.