Re: Portable Assembly

From	Don Y <blockedofcourse@foo.invalid>
Newsgroups	comp.arch.embedded
Subject	Re: Portable Assembly
Date	2017-06-13 23:58 -0700
Organization	A noiseless patient Spider
Message-ID	<ohqmk3$n9d$1@dont-email.me> (permalink)
References	(13 earlier) <o2kujc1a2deaibdv40ilpc8seanjdn0pok@4ax.com> <ohntot$6ec$1@dont-email.me> <1a6033ff-f90a-4fd8-8f78-99b83fccab9a@googlegroups.com> <ohqfgq$5fo$1@dont-email.me> <ohqh84$9g4$1@dont-email.me>

Show all headers | View raw

On 6/13/2017 10:26 PM, Reinhardt Behm wrote:
> AT Wednesday 14 June 2017 12:57, Don Y wrote:
>
>> Correct.  But, it's only "order of" assessments.  I.e., is this
>> a constant time algorithm?  Linear time?  Quadratic?  Exponential?
>> etc.
>>
>> There's a lot of handwaving in O() evaluations of algorithms.
>> What's the relative cost of multiplication vs. addition operators?
>> Division?  etc.
>
> I found many of such evaluation grossly wrong. We had some iterative
> solutions praised as taking much fewer iterative steps than others. But
> nobody took into account that each step was much more complicated and took
> much more CPU time, than one of the solutions that took more steps.
> And quite often the simpler solution worked for the more general case
> whereas the "better" worked only under limited conditions.

That doesn't invalidate the *idea* of modeling algorithms using some
set of "abstract operation costs".

But, it points to the fact that the real world eventually intervenes;
and, its a "bitch".  :>  Approximations have to eventually give way to
REAL data!

For example, I rely heavily on the (paged) MMU in the way my RTOS
handles "memory objects".  I'll map a page into another processes
address space instead of bcopy()-ing the contents of the page
across the protection boundary.  A win, right?  "Constant time"
operation (instead of a linear time operation governed by the
amount of data to be bcopied).

But, there's a boat-load of overhead in remapping pages -- including
a trap to the OS to actually do the work.  So, what you'd *think* to be
a more efficient way of handling the data actually is *less* efficient.

(Unless, of course, you can escalate the amount of data involved
to help absorb the cost of that overhead)

OTOH, there are often "other concerns" that can bias an implementation
in favor of (what appears to be) a less efficient solution.

E.g., as I work in a true multiprocessing environment, I want to
ensure a caller can't muck with an argument for which it has provided
a "reference" to the called function WHILE the called function is
(potentially) using it.  So, the MMU lets me mark the page as
immutable (even though it *should* be mutable) UNTIL the called
function has completed.

[This allows the caller to *access* the contents of the page
concurrently with the called function -- but, lets the OS
intervene if the caller tries to alter the contents prematurely.
The "write lock" has value that couldn't be implemented with the
bcopy() approach -- think of large objects -- and is "cheaper"
to implement once you've already decided to take the hit for
manipulating the page tables]

The "real machine" fixes values to all those constants (K) in the
O()-evaluation.  And, can muck with the decision making process
in certain sets of constraints.

Thread

Portable Assembly rickman <gnuarm@gmail.com> - 2017-05-27 15:39 -0400
  Re: Portable Assembly Grant Edwards <invalid@invalid.invalid> - 2017-05-27 21:05 +0000
    Re: Portable Assembly Mel Wilson <mwilson@the-wire.com> - 2017-05-27 21:29 +0000
    Re: Portable Assembly upsidedown@downunder.com - 2017-05-28 00:36 +0300
      Re: Portable Assembly Grant Edwards <invalid@invalid.invalid> - 2017-05-27 21:56 +0000
        Re: Portable Assembly Hans-Bernhard Bröker <HBBroeker@t-online.de> - 2017-05-28 14:36 +0200
      Re: Portable Assembly David Brown <david.brown@hesbynett.no> - 2017-05-29 09:53 +0200
  Re: Portable Assembly Les Cargill <lcargill99@comcast.com> - 2017-05-27 16:17 -0500
    Re: Portable Assembly Dimiter_Popoff <dp@tgi-sci.com> - 2017-05-28 00:31 +0300
      Re: Portable Assembly Theo Markettos <theom+news@chiark.greenend.org.uk> - 2017-05-27 22:52 +0100
        Re: Portable Assembly Grant Edwards <invalid@invalid.invalid> - 2017-05-27 21:58 +0000
          Re: Portable Assembly upsidedown@downunder.com - 2017-05-28 07:37 +0300
            Re: Portable Assembly rickman <gnuarm@gmail.com> - 2017-05-28 01:28 -0400
            Re: Portable Assembly Tom Gardner <spamjunk@blueyonder.co.uk> - 2017-05-28 07:33 +0100
              Re: Portable Assembly rickman <gnuarm@gmail.com> - 2017-05-28 03:22 -0400
                Re: Portable Assembly Tom Gardner <spamjunk@blueyonder.co.uk> - 2017-05-28 08:49 +0100
                Re: Portable Assembly upsidedown@downunder.com - 2017-05-28 15:03 +0300
          Re: Portable Assembly Tauno Voipio <tauno.voipio@notused.fi.invalid> - 2017-05-28 12:26 +0300
        Re: Portable Assembly Dimiter_Popoff <dp@tgi-sci.com> - 2017-05-28 01:04 +0300
        Re: Portable Assembly David Brown <david.brown@hesbynett.no> - 2017-05-29 10:43 +0200
          Re: Portable Assembly Anssi Saari <as@sci.fi> - 2017-05-31 10:19 +0300
            Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-05-31 01:49 -0700
            Re: Portable Assembly Tauno Voipio <tauno.voipio@notused.fi.invalid> - 2017-05-31 15:56 +0300
        Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-05-29 02:33 -0700
          Re: Portable Assembly upsidedown@downunder.com - 2017-05-29 23:46 +0300
            Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-05-29 13:54 -0700
      Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-05-27 15:52 -0700
        Re: Portable Assembly Dimiter_Popoff <dp@tgi-sci.com> - 2017-05-28 02:14 +0300
          Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-05-27 17:14 -0700
            Re: Portable Assembly Dimiter_Popoff <dp@tgi-sci.com> - 2017-05-28 03:38 +0300
              Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-05-27 18:42 -0700
      Re: Portable Assembly Stefan Reuther <stefan.news@arcor.de> - 2017-05-28 18:45 +0200
        Re: Portable Assembly Dimiter_Popoff <dp@tgi-sci.com> - 2017-05-28 20:47 +0300
          Re: Portable Assembly Stefan Reuther <stefan.news@arcor.de> - 2017-05-29 18:43 +0200
            Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-05-29 10:23 -0700
            Re: Portable Assembly "Boudewijn Dijkstra" <sp4mtr4p.boudewijn@indes.com> - 2017-05-30 16:31 +0200
        Re: Portable Assembly Grant Edwards <invalid@invalid.invalid> - 2017-05-29 01:59 +0000
          Re: Portable Assembly Dimiter_Popoff <dp@tgi-sci.com> - 2017-05-29 05:21 +0300
      Re: Portable Assembly Les Cargill <lcargill99@comcast.com> - 2017-05-28 22:04 -0500
        Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-05-29 02:57 -0700
      Re: Portable Assembly David Brown <david.brown@hesbynett.no> - 2017-05-29 11:00 +0200
        Re: Portable Assembly Dimiter_Popoff <dp@tgi-sci.com> - 2017-05-29 15:08 +0300
          Re: Portable Assembly David Brown <david.brown@hesbynett.no> - 2017-05-29 15:06 +0200
            Re: Portable Assembly Dimiter_Popoff <dp@tgi-sci.com> - 2017-05-29 20:02 +0300
              Re: Portable Assembly David Brown <david.brown@hesbynett.no> - 2017-05-29 23:13 +0200
                Re: Portable Assembly Dimiter_Popoff <dp@tgi-sci.com> - 2017-05-30 16:53 +0300
                Re: Portable Assembly David Brown <david.brown@hesbynett.no> - 2017-05-31 11:36 +0200
                Re: Portable Assembly Dimiter_Popoff <dp@tgi-sci.com> - 2017-06-01 22:43 +0300
                Re: Portable Assembly David Brown <david.brown@hesbynett.no> - 2017-06-02 13:27 +0200
          Re: Portable Assembly Grant Edwards <invalid@invalid.invalid> - 2017-05-29 13:48 +0000
    Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-05-27 15:46 -0700
    Re: Portable Assembly Tim Wescott <tim@seemywebsite.really> - 2017-05-28 23:59 -0500
  Re: Portable Assembly Dimiter_Popoff <dp@tgi-sci.com> - 2017-05-28 00:23 +0300
  Re: Portable Assembly jim.brakefield@ieee.org - 2017-05-27 16:54 -0700
  Re: Portable Assembly Walter Banks <walter@bytecraft.com> - 2017-05-29 17:23 -0400
  Re: Portable Assembly "Boudewijn Dijkstra" <sp4mtr4p.boudewijn@indes.com> - 2017-06-02 17:03 +0200
    Re: Portable Assembly Mike Perkins <spam@spam.com> - 2017-06-05 15:39 +0100
      Re: Portable Assembly David Brown <david.brown@hesbynett.no> - 2017-06-05 18:10 +0200
      Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-06-05 09:10 -0700
        Re: Portable Assembly George Neuner <gneuner2@comcast.net> - 2017-06-06 00:24 -0400
          Re: Portable Assembly David Brown <david.brown@hesbynett.no> - 2017-06-06 11:05 +0200
          Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-06-06 07:03 -0700
            Re: Portable Assembly David Brown <david.brown@hesbynett.no> - 2017-06-06 16:58 +0200
              Re: Portable Assembly Tauno Voipio <tauno.voipio@notused.fi.invalid> - 2017-06-06 21:26 +0300
            Re: Portable Assembly George Neuner <gneuner2@comcast.net> - 2017-06-06 16:42 -0400
              Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-06-06 23:16 -0700
                Re: Portable Assembly George Neuner <gneuner2@comcast.net> - 2017-06-08 06:38 -0400
                Re: Portable Assembly Dimiter_Popoff <dp@tgi-sci.com> - 2017-06-08 15:58 +0300
                Re: Portable Assembly upsidedown@downunder.com - 2017-06-08 21:34 +0300
                Re: Portable Assembly Les Cargill <lcargill99@comcast.com> - 2017-06-10 15:27 -0500
                Re: Portable Assembly Les Cargill <lcargill99@comcast.com> - 2017-06-10 15:23 -0500
                Re: Portable Assembly Tom Gardner <spamjunk@blueyonder.co.uk> - 2017-06-08 14:50 +0100
                Re: Portable Assembly Tauno Voipio <tauno.voipio@notused.fi.invalid> - 2017-06-08 17:38 +0300
                Re: Portable Assembly Tom Gardner <spamjunk@blueyonder.co.uk> - 2017-06-08 16:37 +0100
                Re: Portable Assembly George Neuner <gneuner2@comcast.net> - 2017-06-09 14:14 -0400
                Re: Portable Assembly Tom Gardner <spamjunk@blueyonder.co.uk> - 2017-06-09 22:25 +0100
                Re: Portable Assembly Clifford Heath <no.spam@please.net> - 2017-06-10 09:54 +1000
                Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-06-09 00:06 -0700
                Re: Portable Assembly George Neuner <gneuner2@comcast.net> - 2017-06-09 22:14 -0400
                Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-06-09 22:50 -0700
                Re: Portable Assembly George Neuner <gneuner2@comcast.net> - 2017-06-10 21:01 -0400
                Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-06-11 00:39 -0700
                Re: Portable Assembly George Neuner <gneuner2@comcast.net> - 2017-06-12 23:27 -0400
                Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-06-12 22:42 -0700
                Re: Portable Assembly Ed Prochak <edprochak@gmail.com> - 2017-06-13 14:24 -0700
                Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-06-13 21:57 -0700
                Re: Portable Assembly Reinhardt Behm <rbehm@hushmail.com> - 2017-06-14 13:26 +0800
                Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-06-13 23:58 -0700
                Re: Portable Assembly Ed Prochak <edprochak@gmail.com> - 2017-06-19 06:48 -0700
                Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-06-19 16:45 -0700
                Re: Portable Assembly Ed Prochak <edprochak@gmail.com> - 2017-06-20 15:07 -0700
                Re: Portable Assembly Don Y <blockedofcourse@foo.invalid> - 2017-06-21 07:37 -0700
              Re: Portable Assembly David Brown <david.brown@hesbynett.no> - 2017-06-07 09:20 +0200
          Re: Portable Assembly Niklas Holsti <niklas.holsti@tidorum.invalid> - 2017-06-07 00:30 +0300
            Re: Portable Assembly George Neuner <gneuner2@comcast.net> - 2017-06-07 02:02 -0400

csiph-web