Path: utzoo!utgpu!news-server.csri.toronto.edu!clyde.concordia.ca!uunet!wuarchive!sdd.hp.com!uakari.primate.wisc.edu!unmvax!ariel.unm.edu!triton.unm.edu!keirje
From: keirje@triton.unm.edu (Keirje Aldranske)
Newsgroups: comp.os.msdos.programmer
Subject: Re: realtime CPU speed (effective or actual)
Message-ID: <1990Oct15.043746.23946@ariel.unm.edu>
Date: 15 Oct 90 04:37:46 GMT
References: <1990Oct14.192245.15414@ariel.unm.edu> <27191cec.5488@petunia.CalPoly.EDU>
Sender: news@ariel.unm.edu (USENET News System)
Organization: University of New Mexico, Albuquerque NM
Lines: 60

In article <27191cec.5488@petunia.CalPoly.EDU> jdudeck@polyslo.CalPoly.EDU (John R. Dudeck) writes:
>
>
>In an article keirje@triton.unm.edu (Keirje Aldranske) wrote:
>>  How does one determine the *effective* or even the actuall cpu speed
>>  in realtime? 
>>
>>  I can run a timed loop and find load average (X Windows Style) but I want
>>  the effective mhz.  I know there are several programs which do this:
>>  MicroScope(PC Mag) and CPUSPEED and a few other... any help will be
>>  appreciated.
>
>When you say "effective cpu speed", I question what that is supposed to
>mean.  After all, even when the system is sitting in COMMAND.COM
>running in a wait loop, it is running at the full CPU speed.
>
>Are you thinking of some sort of tachometer which would indicate what
>amount of useful work is being done, referenced to the CPU clock speed?
>For example, if my clock speed is 25 Mhz, and I am sitting in a wait
>loop, it should indicate 0 Mhz, and if it is doing full number crunching
>to recalculate a spreadsheet it should read 25 Mhz.  During disk access
>it should indicate something in between, etc?
>
>The only way that would be possible is if DOS had some sort of system
>state variable that indicates what it is doing, and which could be
>monitored by a TSR program, which would then calculate the effective
>speed and display it.  This is what I believe you are suggesting by 
>displaying the "load average", but as far as I know this is not possible
>because this information does not exist in DOS.
>
>But perhaps you want a program that measures the CPU clock frequency,
>working backwards from the known number of CPU cycles necessary to
>perform some operation, and timing the rate at which the operation is
>executed.  This would be what you mean by "actual CPU speed".  There
>are programs that do this, though I can't remember what any of them are
>called, or where to find them.
>
>But there is another type of information that is useful to know.  This
>is the percentage of time the CPU is not doing useful work, such as
>during RAM refresh, RAM wait states, I/O bus wait states, etc.  Also it
>would be interesting to know how much time is spent waiting for disk
>I/O.  This is the type of information that is compared in benchmarks,
>though I've never seen one which displays it in this way.  Usually they
>give a figure of merit that shows the overall effect of these elements
>combined with the CPU speed, and resulting in a relative, not absolute
>performance figure.  Norton's SI would be in this category.
>
>-- 
>John Dudeck                                  "Nothing is foolproof, because
>jdudeck@Polyslo.CalPoly.Edu                        fools are so ingenious."
>ESL: 62013975 Tel: 805-545-9549                       -- quote from PC Mag.



   I am talking about running this program on a multi-tasking machine.  
   ie. DesqView or Windows 3.0  I already have a working version of
   an XLoad type program.  I want to be able to determine effective CPU
   mhz as you mentioned in your posting for a tachometer.

            - Kay