Computer Chess Club Archives




Subject: Re: assembly--not really that fast

Author: Ed Schröder

Date: 01:05:25 01/15/02

Go up one level in this thread

On January 14, 2002 at 23:25:13, Jonathan Parle wrote:

>On January 14, 2002 at 04:32:32, Ed Schröder wrote:
>>On January 14, 2002 at 03:26:56, Jonathan Parle wrote:
>>>On January 14, 2002 at 01:04:37, Uri Blass wrote:
>>>>On January 13, 2002 at 23:11:50, Jonathan Parle wrote:
>>>>>Your response helps explain (at least to me) why I have specifically always
>>>>>found your programs to be the toughest of all to play against (along with
>>>>>Richard Lang's) at fast time controls. I notice this advantage tends to
>>>>>dissipate with increased time controls - to the point where a lot of the
>>>>>programs seem to me to be fairly similar in strength - just with differing
>>>>>playing styles.
>>>>There are better programs in comp-comp games also in blitz.
>>>>Genius was the best some years ago but it is not the best today even not in
>>>>Century4 is better than Genius but is not known to be number 1 in blitz.
>>>>I cannot explain why do you find Century and Genius better than other programs
>>>>if you have the other top programs of today.
>>>I was referring to my own games against the programs rather than comp vs comp.
>>>So I would attribute this to my playing style. I concur with what you say
>>>regarding comp vs comp. However, just because player A, B and C share a similar
>>>rating, doesn't prevent player C from walloping player A in a 10 game match and
>>>player C getting walloped by player B in a ten game match. Some players of
>>>similar rating simply find they have difficutly against particular opponents of
>>>a similar rating, but not others. I think this is what is happening here with my
>>>human vs comp experiences. I just find Ed Shroeder's and Richard Lang's programs
>>>much tougher blitz opponents than, say Junior or Fritz, regardless of what the
>>>published ratings say.
>>I often get those remarks, for instance: Rebel so now and then (but more than
>>other programs) can play moves that at first sight look as a major blunder. It's
>>no blunder, it's a trap, if you bite you are toasted. Naturally in comp-comp
>>this is worthless, it only counts in human-comp.
>>Genius: a spider in his web, waiting for a tiny mistake and then strike. I can
>>tell, Rebel lost many major games against the spider in the old 6502 days.
>Ed, without me getting too nostalgic and teary eyed, our family's most prized
>chess-related possesion is actually our Mephisto Exclusive MMIV, bought brand
>new in 1988. Boy, you sure were the king of 6502 assembly programming. How on
>earth you got such unbelievable playing strength out of a 5Mhz 8 bit 6502 I'll
>never know.

[ In nostalgic mode ]

Frans Morsch in those days did an even better job, 4 Mhz 6502, 16 Kb rom, 512
bytes ram (the Mephisto Mondial) and it was very close in strength to my stuff
wheras my stuff ran on 5 Mhz, and 4 Kb ram (Mephisto Rebel). Don't ask me how
Frans did it, but he did.

In those days there wasn't any C compiler so you were forced to write in
assembler. Writing assembler it was not the chess engine that gave me a headache
(I was used to that delicate needle work) but writing the interface was a major
pain for me each time. You have to realize you had write directly into the
hardware, make those LED's burn, program the buzzer, write to the screen (you
had to program the characters yourself), program the keys all driven my some
tricky interrupt routine. Bah, I never liked it.

Later the hardware was upgraded to 32 Kb rom and 8 Kb ram, it gave some extra
elo but not all that exciting. At a certain moment you realize the hardware
limitation is a dead-end street, time to say goodbye to the 6502 and move on to
the RISC processor (the ChessMachine) and a few years later to the PC.

[ back to real life, wiping a tear ]


>I remember playing blitz against the demo version of your MMIV at the shop the
>day we made the purchase, as well as against the then new Mephisto Almeria 16
>bit. Not that we could afford the Almeria, I just wanted see how much stronger
>it was than your program. Well I lost 5 out of 5 games of blitz to your MMIV,
>but won 2 out of 5 against Almeria.
>Interestingly, last year I played a 12 game match between the Exclusive MMIV and
>my Mephisto Miami. The Miami won by a narrow margin of 6.5 - 5.5. This was at a
>time control of 40 in 2. At anything below 30 secs a move, the MMIV would
>probably win every game. It was an interesting square-off for me, because on the
>one hand I had a machine with a 32K program running on a 5Mhz 6205 versus a 16K
>program running on a 10Mhz RISC processor.
>I reckon the golden days were 1982 to 1989. Maybe that's because during that
>period, dedicated machines well and truly exceeded me in playing ability, as was
>the case for 99% of chess players. It would be really nice to see you still do a
>program for a dedicated machine every now and then, but it looks like Frans
>Morsch has that market sewn up??? I know most people on this board are more
>interested in the latest software, but for me there will never be a substitute
>for the good old dedicated machines.

This page took 0.03 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

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