Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Assembly in VC6.0 – an (bad) experience!

Author: Bo Persson

Date: 10:30:43 03/15/01

Go up one level in this thread


On March 15, 2001 at 09:57:54, James Swafford wrote:

>On March 15, 2001 at 02:58:48, José Carlos wrote:
>
>>
>>  I haven't written assembler pieces in my program, so I cannot say about how
>>much can assembler speed up the program, but I think you should feel happy if
>>the compiler does so good job that you don't need to write assembler code,
>>because this way, you have a more portable and understandable (easy to modify)
>>code. Don't you think?
>
>While that's true, on some level I'm a little disappointed.  Assembly
>programming is almost an art unto itself;  now it seems the advantage
>of using assembly may be dwindling.

Yes, but in part this is because the processor architectures (from Intel, at
least) are aimed at what an optimizing compiler can do. For the Itanium, it is
really part of the design that a compiler can schedule the instruction triples
(3 instructions per clock) to take advantage of the parallellism. Even if you
are a very good assembly programmer, you will eventually give up when one
changed instruction forces you to reschedule the rest of the function. A
compiler does this everytime, without getting tired :-)

>>
>>  José C.


Bo Persson
bop@malmo.mail.telia.com



This page took 0.02 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.