Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: I need advise !

Author: Roberto Waldteufel

Date: 11:59:29 04/19/99

Go up one level in this thread


On April 19, 1999 at 14:14:37, James Robertson wrote:

>
>>I must disagree with you. Firstly, speed is critical for chess. How to get
>>maximum speed? - use hand-optimized assembler. That is NOT lunacy, it is plain
>>common sense. It might be difficult, but if the results are better it certainly
>>does not make it mad. Much of my program is written in assembler, so I shall be
>>expecting a visit from the men in the long white coats soon!
>
>You miss my point. Obviously, it cannot be complete lunacy, as Franz Morsch
>actually did it, and he has one of the most successful programs in the world. I
>was talking tongue in cheek. Several speed-critical functions in my program are
>written in hand-coded assembler too.
>
>But, whatever it is, it cannot be common sense. As far as I know, Fritz and
>Rebel are the only PC programs written in assembler. Both are commercial, and
>apparently, Ed Schroeder had enough of assembler, as he announced he was
>translating his program into C++ (very wise, if you ask me).
>
Well, I guess what I mean is "if you can do it, then it makes sense to". It is
the perfectionist's choice - if you refuse to compromise efficiency under any
circumstances, then assembler it has to be. Obviously it is much easier to use a
high level language, but the easiest way does not produce the absolute best
performance. It is no surprise to me that the existing assembler chess programs
are among the best.

>>
>>Second, Visual Basic is more than 20 times slower than the best commercial >Basic
>>compilers
>
>Yes, but I wasn't talking about Basic; I was talking about VB.
>
>>- if you are going to talk about Basic, at least see what a modern
>>efficient (as opposed to virtual anything!) Basic compiler can do. There are
>>Basic compilers now that will compile small, fast executables to run under Win9x
>>and Win NT, and they include a full 32-bit in-line assembler as well - blows the
>>wheels of Visul C, and C++.
>
>The Basic compilers are faster than VC++? I haven't heard this before; perhaps
>you could give me some websites for more information? I am also a good Basic
>programmer, and if it is faster.....
>
>James
>

Yes, it is faster, and you are not alone in not having heard. Check out the
PowerBasic web site at http://powerbasic.com/
For 32-bit Windows stuff, you will want either PBDLL or PBCC compilers (I use
both). In my experience, their products are very good. There are some
discussions about PowerBasic speed compared to other compilers on the BBS there
- see for yourself. If you like the Basic language, then these 32-bit Windows
compilers are really great. PBDLL is more general in its use, whereas PBCC is
much easier to use and very much more similar in its syntax to traditional
basics than VB, but only compiles console (ie text) applications, although I did
manage to design a passable chess display with it. With PBDLL you get the full
graphics capabilities of Windows, but you need to know a lot more about the
Windows API.

The general consensus among programmers who use these compilers as well as other
compilers (like VB, Virtual C and C++) is that PowerBasic is faster, especially
compared to the crawling VB. Compared to straight C the difference is only very
small, but seems to be in Basic's favour. Many of these programmers are people
who use PB at home, but are obliged to use something else at work.

Best wishes,
Roberto



This page took 0 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.