CU Amiga Super CD-ROM 17/CU Amiga Magazine's Super CD-ROM 17 (1997)(EMAP Images)(GB)[!][issue 1997-12].iso/CUCD/Programming/AMOSList/AMOSLIST.0897/[e-mail]_Tue Aug 12 06:18:18 1997.msg

Text File | 1997-09-09 | 4KB | 93 lines

Received: from svcs1.digex.net (svcs1.digex.net [204.91.197.224])
    by mail1.access.digex.net (8.8.5/8.8.5) with ESMTP id GAA21818
    for <[e-mail]>; Tue, 12 Aug 1997 06:18:16 -0400 (EDT)
Received: (from daemon@localhost)
    by svcs1.digex.net (8.8.5/8.8.5) id FAA03575
    for amos-out; Tue, 12 Aug 1997 05:01:19 -0400 (EDT)
Received: from mail3.access.digex.net (mail3.access.digex.net [205.197.247.4])
    by svcs1.digex.net (8.8.5/8.8.5) with ESMTP id FAA03572
    for <[e-mail]>; Tue, 12 Aug 1997 05:01:18 -0400 (EDT)
Received: from MCSDEV.MCS.CO.UK ([e-mail] [193.32.50.112])
    by mail3.access.digex.net (8.8.5/8.8.5) with SMTP id FAA22593
    for <[e-mail]>; Tue, 12 Aug 1997 05:01:02 -0400 (EDT)
Date:     Tue, 12 Aug 1997 10:01 GMT
From: "SIMONC%MARS.decnet"@mcs.co.uk (Simon Champion - [e-mail])
Message-Id: <[e-mail]>
To: [e-mail]
Subject:  Re: Why is Amos SLower than C/C++
X-VMS-To: MCSDEV::SMTP%"[e-mail]"
Status: O
X-Status:

Subj:    Why is Amos SLower than C/C++

Scott Matott wrote:

 >    Quick Question, why is complied AMos Code slower than compiled
 >C/C++ code?  Maybe it's not.  I could see why Assembler would be quicker
 >but why C, is it because better faster libraries are available? Or is it
 >that C compilers are better optimisers? Or is C jsut faster doing some
 >stuff because you can declare you variable size (ie work with Char's
 >instead of Long's)  Well I was jsut wondering if anybody out there knows
 >anything about this as I might go buy a C compiler to write my Caster
 >since no matter what I do it seems slow in Amos.

The basic reason why Amos is slower than C is "Basic".

{lecture mode=on}

Basic is a higher level language than C. Translated into English, this
means that the code is easier for a human to understand - and therefore
harder for a computer to understand - than C code.
When you compile the code, a line of C code is translated into fewer machine
code instructions than a similar Basic instruction.

Let me think of an example......

Consider a simple loop:
   Basic (ie Amos)               C

   For X=1 to 10              while(x=1,x++,x>=10) {
      Print X                    printf("%d",x)
   Next                       }

(You'll excuse any dumb mistakes - I've not done any C code for a while :-)
but it's good enough for the example)

These two chunks of code do the same thing (roughly :), but after compiling,
the C code is much smaller.

In Basic, the Print statement gets converted to:
   1. Make a string version of variable X.
   2. Append this string X to the rest of the string (nothing in this case)
   3. Print the string.
In C, the code would be compiled as:
   1. Insert the decimal value of x into the string
   2. Print the string.

(this is just an example, so please don't jump on me if that isn't exact :)

The compiled Basic has one more thing to do, and that also happens to be
the slowest step in the process. The loop instructions also have the same
sorts of differences, as do large chunks of all Basic programs.

The more complex commands in Basic, and Amos especially, can blossom into
large chunks of compiled code, and unless you use every available feature
of the command every time you call it, there is bound to be some wastage
in the compiled code (like the second step of the Print example), which
could often be avoided when using C, and always avoided when using Asm.
(assuming, of course, you know C or asm.... :)

{lecture mode=off}  :-)

Ok. Hope that helped........ (hope it didn't bore you witless)...... :))

Seeyall!

    / |\   /
    \ | \ /
    / |_/ \_
 (Simon Champion)    <[e-mail]>

-------------------------------------------------------------------------------
"Here speeking American Enlish well talking"  - Sign in Majorca shop.