HP3000-L Archives

January 1996, Week 2

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Bruce Toback <[log in to unmask]>
Reply To:
Bruce Toback <[log in to unmask]>
Date:
Thu, 11 Jan 1996 08:52:30 -0700
Content-Type:
text/plain
Parts/Attachments:
text/plain (29 lines)
Bill Lancaster writes:
>Now that we have covered many of the syntactical issues about the Multiple
>Job Queues, I thought that I would ask the list what they thought about the
>idea of being able to segregate a particular processor in a multiple
>processor environment to a particular queue/workgroup?
 
I'd argue against this because it's at once too little control and too much
control. The Workload Manager allocates total processing resources to
different workloads, regardless of the number of processors. You can
achieve the same effect by telling the Workload Manager to allocate 1/4th
of the available CPU time on a 4-processor system to a given workload. But
whether that ends up being one processor all the time or all four
processors 1/4th of the time shouldn't make any difference. There's no
point in having a processor idle just because there's no work at any given
instant for its workgroup.
 
The only possible advantage to this would be to increase cache locality,
and it's doubtful whether the resulting performance improvement would be
measurable in a typical application.
 
-- Bruce
 
---------------------------------------------------------------------------
Bruce Toback    Tel: (602) 996-8601| My candle burns at both ends;
OPT, Inc.            (800) 858-4507| It will not last the night;
11801 N. Tatum Blvd. Ste. 142      | But ah, my foes, and oh, my friends -
Phoenix AZ 85028                   | It gives a lovely light.
[log in to unmask]                 |     -- Edna St. Vincent Millay

ATOM RSS1 RSS2