Simple Question about nodes in dedicated chess computers

This forum is for general discussions and questions, including Collectors Corner and anything to do with Computer chess.

Moderators: Harvey Williamson, Steve B, Watchman

Forum rules
This textbox is used to restore diagrams posted with the fen tag before the upgrade.
Post Reply
User avatar
appleshampogal
Member
Posts: 126
Joined: Sun Jan 12, 2014 8:53 am

Simple Question about nodes in dedicated chess computers

Post by appleshampogal »

In my dedicated computer, mephisto chess challenger, are the nodes expressed (which on average around 3,000) expressed as kilonodes? Or does 3,000 nodes mean strictly three thousand potential future board positions a second? I figure it could be kilonodes considering something I read here equating my unit with the power of a 386x.

Thoughts anyone?[/list]
Larry
Senior Member
Posts: 2269
Joined: Wed Aug 01, 2007 8:42 am
Location: Gosford, NSW Australia

Post by Larry »

3000 nodes per second only. That's why we have to take our hats off
to those early programmers.
L
User avatar
appleshampogal
Member
Posts: 126
Joined: Sun Jan 12, 2014 8:53 am

Post by appleshampogal »

Ohhhh, well alrighty then. Jeez, it's amazing how different numbers are nowaday's.

Do nodes work as stated, ie looking at thousands of future board positions that could result from the projected outcome the computer thinks will come to pass? I imagine due to the level of selective search that this (and other) dedicated units must have is somewhat of a preemptive compensation for the low node count? In other words, by making a program more selective, it would allow for the program to search more deeply into lines it finds most promising due to the selective approach thereby mitigating it's otherwise slower speed?
Larry
Senior Member
Posts: 2269
Joined: Wed Aug 01, 2007 8:42 am
Location: Gosford, NSW Australia

Post by Larry »

appleshampogal wrote:
Do nodes work as stated, ie looking at thousands of future board positions that could result from the projected outcome the computer thinks will come to pass? I imagine due to the level of selective search that this (and other) dedicated units must have is somewhat of a preemptive compensation for the low node count? In other words, by making a program more selective, it would allow for the program to search more deeply into lines it finds most promising due to the selective approach thereby mitigating it's otherwise slower speed?
Hi, I don't pretend to be some sort of guru on this matter. For good
information go to: http://talkchess.com/forum/index.php
I can tell you that the earlier programmers had to juggle selectivity with
full width searches. For a feeble processor, selectivity will increase the
chance of a good line being found but also increase the chance of committing itself to a bad line. So you understand that a playing strength
in the high teens at 3000 nodes/sec was an achievement, especially considering they did not
even have hash tables to speed up the search.
L
User avatar
Cyberchess
Full Member
Posts: 658
Joined: Wed Jan 08, 2014 6:10 pm

Post by Cyberchess »

Back in the personal computing heyday of the Intel 386 with 32 bit architecture, the Windows 3.1 OS was en vogue as this allowed non computer savvy folks to point and click rather than enter command line instructions. This early version of Windows was extremely bug ridden, patched many times over, and unfortunately negated much of the 386’s potential. These machines struggled to perform basic tasks such as word processing, and their chess playing strength suffered accordingly.

Rather than subject my PC to this unnecessary abuse, I used to run everything through a batch driven DOS menu system. By maximizing the memory, running lean and using a decent chess program, my old 386 came to be my preferred chess partner of the early ‘90s. It wouldn’t be until 1995 that I would purchase another dedicated unit (the Novag Diamond). We used to periodically check the results published in the excellent publication “Selective Search” by Eric Hallsworth.

http://www.elhchess.demon.co.uk/sscontnt.htm
User avatar
appleshampogal
Member
Posts: 126
Joined: Sun Jan 12, 2014 8:53 am

Post by appleshampogal »

I remember the DOS era fondly. Thank you for that link as well. Now I know the top dedicated units! Milano pro/Mephisto Master is number 10! I saw one on eBay that is selling for $179 or so. If anyone can tell me if that is a good price let me know. I'm definitely interested in the technical aspects of how my dedicated plays. I'll try searching for my machine and see what comes up. If I get the milano pro it might be fun to post a game against the challenger, just for giggles. From what I understand, the colditz test gave the challenger a 2,021 and the milano a 2,156. Given what I understand about elo, that's quite the difference. Wonder how it'll play out.
User avatar
Cyberchess
Full Member
Posts: 658
Joined: Wed Jan 08, 2014 6:10 pm

Post by Cyberchess »

Glad to hear that you found Selective Search to be a useful reference. I seem to recall mention of the Colditz Test many years ago, but don’t remember the specifics. It was likely something that I read about in Chess Life way back when.

I do recall also seeing the Mephisto Master Chess Computer selling for $179 on the bay recently. Do you also own the Saitek Chess Challenger unit? A chess playing friend recently listed his hardly used Novag Sapphire portable unit with a Buy It Now price of $160. He said that it sold in under an hour.

Computer matches are always entertaining, though I never had the patience or enough spare time to have them play at tournament time. I used to pair them at Sudden Death 60 in order to expedite the match.
User avatar
appleshampogal
Member
Posts: 126
Joined: Sun Jan 12, 2014 8:53 am

Post by appleshampogal »

Yes I do! I nicknamed my Mephisto Chess Challenger, "Clementine". I saw one ad for the Mephisto Milano Pro, but it was German and the ad stated that there was potential corrosion to the battery area, yikes.

The thing I like about computer-computer matches is that sometimes it's difficult to figure out what went wrong in the losing side. Obviously there are clear times where it is, but not always. Clementine seems to have a frustrating defensive style, but I hear the master milano is more aggressive, which should make for an interesting collection piece.
User avatar
Cyberchess
Full Member
Posts: 658
Joined: Wed Jan 08, 2014 6:10 pm

Post by Cyberchess »

Clementine I believe is one of a multitude of Franz Morsch authored programs running on an H8 Renesas (Hitachi) processor @ 10MHZ. While opening book size, amount of system RAM and features will vary, this combination is employed by my Radio Shack Master Chess unit, The Saitek Centurion, GK 2100, Mephisto MMVI and other machines.
At action chess levels, my RS Master Chess repeatedly drew against the more powerful Novag Diamond with its much larger (36K + User Programmable) opening book. Somehow I think that the Franz Morsch program is more efficient. A more compact program is probably better suited for quick chess.

The Mephisto Milano Pro should be clearly stronger than the aforementioned, although you already own 2 Franz Morsch programmed machines. It may turn out to be strikingly similar to your Mephisto Master Chess unit. B.T.W. – does the Master Chess unit allow you to specify a style such as active, passive, etc.?

So many clones, so little time…

Regards,
John
User avatar
appleshampogal
Member
Posts: 126
Joined: Sun Jan 12, 2014 8:53 am

Post by appleshampogal »

Well actually I don't have the second computer yet. I just created a post seeking one however. I'm not having much luck on eBay...
User avatar
Cyberchess
Full Member
Posts: 658
Joined: Wed Jan 08, 2014 6:10 pm

Post by Cyberchess »

appleshampogal wrote:Well actually I don't have the second computer yet. I just created a post seeking one however. I'm not having much luck on eBay...
Sorry, this was a misunderstanding on my part. I was under the impression that you already owned the MMC unit.

e-Bay can be a real funny venue, with items often selling for far too little or far too much money. Back in the year 2000 I was planning a winter vacation and didn’t yet own a PDA, Pocket PC or smart phone. I wanted to bring along my only portable chess computer at the time, but had misgivings about subjecting my Novag Diamond to the airport baggage circuit. At the time there were a plethora of e-Bay sellers offering new to like-new late ‘90s Saitek/Mephisto made RS units for next to nothing opening bids. I won auctions for 3 such units with the idea of testing them out, bringing the best of the lot on my trip and selling off the remaining 2 units when I got back. I compared the RS 2200X, 2250XL and Master Chess units by playing action chess with them and having them play the Diamond. Without a doubt, the 2250XL was the best looking of the 3, featuring an oversized board and a nifty LCD board besides, though the somewhat bland MC unit was clearly the better player, so this is the one that I kept. I’ve since lugged this machine along on several air and train adventures without any problems. It runs for a very long time on C cells and the pieces stow away in their own compartment, which made for a great portable unit in my pre smart phone days.

Good luck in your search, and don’t go overboard bidding :!:
If you’re patient, eventually another will come along for far less money.

On The Bay Regards,
John
User avatar
appleshampogal
Member
Posts: 126
Joined: Sun Jan 12, 2014 8:53 am

Post by appleshampogal »

The only dedicated chess computer I have is the MCC. I wanted a milano or master to start my collection. I'll then have two as opposed to one. The 2250XL... Isn't that the one that's supposedly a Franz Morsch clone?
User avatar
Cyberchess
Full Member
Posts: 658
Joined: Wed Jan 08, 2014 6:10 pm

Post by Cyberchess »

appleshampogal wrote:The only dedicated chess computer I have is the MCC. I wanted a milano or master to start my collection. I'll then have two as opposed to one. The 2250XL... Isn't that the one that's supposedly a Franz Morsch clone?
Yes, Saitek/Mephisto has rehashed the Franz Morsch code repeatedly over the years. Besides using the same program, the 22550XL also sports a much larger (20K) opening book than that of Clementine, RS MC, Saitek Cosmos, Saitek Centurion (10MHZ), etc. which all come with a rather lean 6K opening book. The major difference as far as I can tell is that the 2250XL -- which was released in 1996 -- uses an earlier generation H8 processor. This earlier version H8 likely executes fewer instructions per second at a given clock speed, and results in the 2250XL not being up to par with the aforementioned machines, despite having a much larger opening book. It’s really too bad because the 2250XL sports a much nicer aesthetic in my opinion.

Chalk it up to time and technology...

Regards,
John
Post Reply