🚀 go-pugleaf

>Try explaining to Assembler programmers though that there will soon
>come a point where optimising compilers will be able to generate *better*
>code than a programmer - they just refuse to believe you.  But the
>increasing advancements in compiler technologies, plus the increasing
>complexity of processors, means that soon generating machine code
>will be a job best left to the computer.  And low level languages
>like Assembler will only be used in a very small handful of situations 
>(if at all !!!).


You should dry some liters of humbleness, once for all.

I'll try to be on your side, let's examinate (and demolish) the best compiler
around: the PC's Watcom C++.

What code Watcom or the best compiler of the 2000s will generate i.e. for a
PlotPixel function? How many x10 times slower than this?

    shl    edi, 6                ; ebx *= 64                    3 clocks
    lea    edi, [edi*4+edi]      ; ebx *= 5   (64*5 = 320)      1 clock
    mov    [edi+ebx+0a0000h], al ;                              memory ref

That does in 3 instructions 1 multiplication (*320), 1 addition (for X),
and the write in the right zone of the screen memory.

Since compilers are dumb and humans (well, except a big part) aren't, one can
also examinate if the cache system is good enough to make become convenient
this 2 instructions version:

    mov    edi, [edi*4 + table]  ;                              memory ref
    mov    [edi+ebx+0a0000h], al ;                              memory ref

---

Now, since me and many others are humans and have a brain that distiguish us
from the machines, that makes us "intelligent" unlike them, I can tell you
that your damn compilers will NEVER generate a better code than a human, that
has the right brain structure to gain from experience and association of
memories/experiences.

Now, if your assembly skills are so crap that an interpreted Basic will be
faster than your assembly code, fair enough.. your problem pal.


I agree only about a thing with you: today and tomorrow compilers will
averagely generate better code than humans, but NOT because compilers are
good; simply because there're a lot of lamers like you going around and
pretending to be called assembly coders.

Well pal, if you feel offended by my reply, remember that 1) I am tired of
reading that assembly coders are imbecills, while they're "ingenious", and
2) that this wanting-to-omologate-everyone sucks bigtime.. I will never be
a sad C++ programmer using sad academical methods and sad tight and senseless
and needless rules.
There're people that are able to make innovative stuffs, and they use and
are able to use assemblers in the proper way.

>>
>>-- 
>>Ben Hutchings, JCR Computer Rep elect, Worcester College, Oxford.

>-FISH-   ><>



  /-----------------------------------------------------------------------\
  |  Fabio "Maverick" Bizzetti - bizzetti@mbox.vol.it - Maverick* at IRC  |
  |            The maker of "CyberMan" and "Virtual Karting"              |
  |              working on "VirtualRally" & "StarFighter"                |
  \-----------------------------------------------------------------------/