Phoenix Chess Systems the makers of the Resurrection and Revelation dedicated chess computers have recently released a Lang Emulation module set for their computers
Running on the Revelation board the clock speed of these emulations are about 4 times the speed of the original modules
so for example..
the Lyon 68000 modules run at 12 Mhz and on the Emulation set about 45 Mhz(including the 25% enhancement ..still in beta testing)
it will be interesting to test the actual speed ups achieved as compared to the original modules on a module by module comparison
http://www.flickr.com/photos/10261668@N ... 195096455/
Bobby Bonds Sends His Regards
Steve
Collector's Corner..Richard Lang On Steroids
Moderators: Harvey Williamson, Steve B, Watchman
Forum rules
This textbox is used to restore diagrams posted with the fen tag before the upgrade.
This textbox is used to restore diagrams posted with the fen tag before the upgrade.
-
Monsieur Plastique
- Senior Member
- Posts: 1014
- Joined: Thu Jul 03, 2008 9:53 am
- Location: On top of a hill in eastern Australia
-
Steve B
- Site Admin
- Posts: 10146
- Joined: Sun Jul 29, 2007 10:02 am
- Location: New York City USofA
- Contact:
Perhaps Sascha(Mythbuster) can answer this more accurately because i know he worked on the emulations with RuudMonsieur Plastique wrote:Were any adjustments made to the hash table sizes from the originals?
my understanding is that if the original program had no hash tables then the speed up is actually larger then if the program used hash tables
so the speed up for the Glasgow -Polgar is on the order of 5x-6x while the speed up for programs using hash tables is on the order of about 4x(12 Mhz to 45 Mhz)
as a result my guess is that the Emulations are using hash tables based upon the Revelations hardware (hash table ram) which would be different then the original programs Hash Ram
Although i could be mistaken about that
Regards
Steve
-
Mythbuster
- Member
- Posts: 275
- Joined: Wed Mar 03, 2010 6:37 pm
Hi,
there are no changes to the original programs. Opening books, RAM, functions: All is 100% original!
Mephisto Amsterdam, Dallas and Roma had no HT
The 68.000 version of Almeria, Portorose, Lyon, Vancouver and London have 512 kb HT
Right now I can say that the emulations are running with the following speed:
Mephisto Amsterdam, Dallas and Roma: 68.000 CPU @ 38 MHz
Mephisto Almeria, Portorose, Lyon, Vancouver and London: 68.000 @ 38 MHz
Mephisto Glasgow: 68.000 @ 66 MHz
Mephisto Rebell 5.0, MM IV, MM V and Polgar: 6502 @ 25 MHz
But this is the actual speed ... Ruud is working on the speed update which will improve the speed for the 68.000 between 25 and 30% ... but this is still in Beta testing ... so please be patient.
Kind Regards,
Sascha
there are no changes to the original programs. Opening books, RAM, functions: All is 100% original!
Mephisto Amsterdam, Dallas and Roma had no HT
The 68.000 version of Almeria, Portorose, Lyon, Vancouver and London have 512 kb HT
Right now I can say that the emulations are running with the following speed:
Mephisto Amsterdam, Dallas and Roma: 68.000 CPU @ 38 MHz
Mephisto Almeria, Portorose, Lyon, Vancouver and London: 68.000 @ 38 MHz
Mephisto Glasgow: 68.000 @ 66 MHz
Mephisto Rebell 5.0, MM IV, MM V and Polgar: 6502 @ 25 MHz
But this is the actual speed ... Ruud is working on the speed update which will improve the speed for the 68.000 between 25 and 30% ... but this is still in Beta testing ... so please be patient.
Kind Regards,
Sascha
-
Monsieur Plastique
- Senior Member
- Posts: 1014
- Joined: Thu Jul 03, 2008 9:53 am
- Location: On top of a hill in eastern Australia
This is going to be a very interesting exercise though I am interested to know why the 68000 versions seem to be emulated and not the 68020 versions?
When I started work on DS Chess, we had an original hash table size of 3072 Kb but after extensive testing we found that the program would sometimes exceed the available memory on the hardware console platform.
We then reduced the hash table size to 2048 Kb and had no further problems. Interestingly, this resulted in a very small speed up for the program - only in the order of about 2% but still measureable. It also had no effect on ply depths reached in the end game, with 22 ply in a King and Pawn test position actually being reached faster than with the bigger tables. This was all running on an ARM9 CPU at 67 Mhz.
One thing that never seems to get measured (at least viewable by the actual user) is actual hash table usage (into total hash allocation). It would be interesting to know this given that when these programs first came out, the hash size was probably more determined by the cost of the memory than it necessarily was the optimal amount. Then again, the Lang program were renowned for their chess knowledge, so maybe like Fruit running on the Nintendo, it did not process enough nodes per second to do much of a dent into large hash table sizes.
Anyway, I look forward to perhaps seeing some Games in 60 - maybe against something like an older version of Fritz or similar running on an early model 90 Mhz Pentium or similar. That would be a clash of two completely different philosophies, where finally the dedicated program gets the hardware to really strut it's stuff.
And as for the non hash programs like Polgar, etc, those Schroder programs were renowned for their performance versus speed of hardware, with improvements de to hardware speed exceeding the average.
I have always felt that a good dedicated from that era could beat something from the early PC engine era (i.e Fritz 5.32 or earlier) given decent hardware, though of course the engines of today are gourged with knowledge, still process nodes very quickly and generally play an extremely intelligent and purposeful game.
Steve, when you finally get your module it might be interesting to have the Polgar play DS Chess at Game in 60. By my figuring the Polgar will be slightly stronger at 25 Mhz but it still could make for an interesting couple of games.
When I started work on DS Chess, we had an original hash table size of 3072 Kb but after extensive testing we found that the program would sometimes exceed the available memory on the hardware console platform.
We then reduced the hash table size to 2048 Kb and had no further problems. Interestingly, this resulted in a very small speed up for the program - only in the order of about 2% but still measureable. It also had no effect on ply depths reached in the end game, with 22 ply in a King and Pawn test position actually being reached faster than with the bigger tables. This was all running on an ARM9 CPU at 67 Mhz.
One thing that never seems to get measured (at least viewable by the actual user) is actual hash table usage (into total hash allocation). It would be interesting to know this given that when these programs first came out, the hash size was probably more determined by the cost of the memory than it necessarily was the optimal amount. Then again, the Lang program were renowned for their chess knowledge, so maybe like Fruit running on the Nintendo, it did not process enough nodes per second to do much of a dent into large hash table sizes.
Anyway, I look forward to perhaps seeing some Games in 60 - maybe against something like an older version of Fritz or similar running on an early model 90 Mhz Pentium or similar. That would be a clash of two completely different philosophies, where finally the dedicated program gets the hardware to really strut it's stuff.
And as for the non hash programs like Polgar, etc, those Schroder programs were renowned for their performance versus speed of hardware, with improvements de to hardware speed exceeding the average.
I have always felt that a good dedicated from that era could beat something from the early PC engine era (i.e Fritz 5.32 or earlier) given decent hardware, though of course the engines of today are gourged with knowledge, still process nodes very quickly and generally play an extremely intelligent and purposeful game.
Steve, when you finally get your module it might be interesting to have the Polgar play DS Chess at Game in 60. By my figuring the Polgar will be slightly stronger at 25 Mhz but it still could make for an interesting couple of games.
Chess is like painting the Mona Lisa whilst walking through a minefield.
-
Steve B
- Site Admin
- Posts: 10146
- Joined: Sun Jul 29, 2007 10:02 am
- Location: New York City USofA
- Contact:
I already have the full emulation package hardwired into my Revelation board
the link i showed in the initial post in this thread is a screen shot of the emulation start-up screen in my Revelation
supposedly the Polgar is running at 25 mhz (5x the normal 5 Mhz)
i also have the "Enhanced" Emulation Suite which is still in beta
it is supposed to speed up the programs an additional 20-25% over the normal emulations
however i am told there is no additional speedup for programs with no hash tables
so the Polgar will still only be running at 25 Mhz even in the Enhanced Emulation suite
Now..Selective Search shows a rating for the Polgar 5 of 1963 and for the Polgar 10 -2038
a 75 pt increase with a 2X speed up
using this as a general guide the Polgar 25 should be around 2115 ish..
I know next to nothing about the "DS"
is this some sort of Cloud Computing chess playing entity like Rybka running on 128 cores or something?
what is the rating of this .."DS"??
Hesitant Regards
Steve
the link i showed in the initial post in this thread is a screen shot of the emulation start-up screen in my Revelation
supposedly the Polgar is running at 25 mhz (5x the normal 5 Mhz)
i also have the "Enhanced" Emulation Suite which is still in beta
it is supposed to speed up the programs an additional 20-25% over the normal emulations
however i am told there is no additional speedup for programs with no hash tables
so the Polgar will still only be running at 25 Mhz even in the Enhanced Emulation suite
Now..Selective Search shows a rating for the Polgar 5 of 1963 and for the Polgar 10 -2038
a 75 pt increase with a 2X speed up
using this as a general guide the Polgar 25 should be around 2115 ish..
I know next to nothing about the "DS"
is this some sort of Cloud Computing chess playing entity like Rybka running on 128 cores or something?
what is the rating of this .."DS"??
Hesitant Regards
Steve
-
Monsieur Plastique
- Senior Member
- Posts: 1014
- Joined: Thu Jul 03, 2008 9:53 am
- Location: On top of a hill in eastern Australia
-
Steve B
- Site Admin
- Posts: 10146
- Joined: Sun Jul 29, 2007 10:02 am
- Location: New York City USofA
- Contact:
Langs London program was the strongest engine ever released for the Mephisto line of computers
The Emulated London contained within the Revelation board runs at about 45 Mhz while the London 030 runs at 33 Mhz
the Mephisto Tournament Machine runs the London at 36 Mhz
In order to play a few games with the Revelation against the TM i would need a desk the size of a small room ..whereas i can cram the London 030 modules into a nice and compact Mephisto modular board
ergo..Perhaps a few ..Revelation London vs Modular London 030 games .. (1 min. Avg./move ) might be of interest
Regards From Big Ben
Steve
The Emulated London contained within the Revelation board runs at about 45 Mhz while the London 030 runs at 33 Mhz
the Mephisto Tournament Machine runs the London at 36 Mhz
In order to play a few games with the Revelation against the TM i would need a desk the size of a small room ..whereas i can cram the London 030 modules into a nice and compact Mephisto modular board
ergo..Perhaps a few ..Revelation London vs Modular London 030 games .. (1 min. Avg./move ) might be of interest
Regards From Big Ben
Steve
-
Monsieur Plastique
- Senior Member
- Posts: 1014
- Joined: Thu Jul 03, 2008 9:53 am
- Location: On top of a hill in eastern Australia
I am a bit surprised - though happy - that you are seemingly all excited about this. I would have thought you might think emulators and running programs at a speed they were not originally designed for was some kind of dedicated chess computer sacrilege.
If you need room for all this stuff I can lease out the spare ballroom in my 19th century mansion if you so desire.
If you need room for all this stuff I can lease out the spare ballroom in my 19th century mansion if you so desire.
Chess is like painting the Mona Lisa whilst walking through a minefield.
-
Steve B
- Site Admin
- Posts: 10146
- Joined: Sun Jul 29, 2007 10:02 am
- Location: New York City USofA
- Contact:
well if they are running on the dreaded PC then of course they are heresyMonsieur Plastique wrote:I am a bit surprised - though happy - that you are seemingly all excited about this. I would have thought you might think emulators and running programs at a speed they were not originally designed for was some kind of dedicated chess computer sacrilege.
not to mention ..Blasphemy even .. because they are not natural
but if the emulators are emulating from a good..ol fashioned..brick and mortar ....work-a-day ..flesh and blood..dedicated computer.. then i welcome them with open arms
Go With The Flow Regards
Steve
-
Steve B
- Site Admin
- Posts: 10146
- Joined: Sun Jul 29, 2007 10:02 am
- Location: New York City USofA
- Contact:
The Passed pawns carry the day for the 030 modules..
Rev-London Resigns here:
[fen]3n1k2/2R5/1p3P2/2p2K1P/P2p4/8/8/8 w - - 0 60[/fen]
Lets Reverse Colors next Regards
Steve
P.S.Full Pgn...
[White "London 68030"]
[Black "Revelation-London"]
[TimeControl "1 Min/Avg."]
[Result "1-0"]
1.e4 Nf6 2.Nc3 e5 3.Nf3 Nc6 4.Bb5 Bb4 5.O-O O-O 6.d3 Bxc3 7.bxc3 d6
8.Bg5 Bd7 9.Nd2 Ne7 10.Bxf6 gxf6 11.Bxd7 Qxd7 12.Qf3 Qe6 13.Rfb1 b6
14.Rb4 f5 15.exf5 Qxf5 16.Qxf5 Nxf5 17.Rg4+ Kh8 18.Ne4 f6 19.Rb1 h5
20.Rg6 d5 21.Nxf6 Ne7 22.Rg5 Rxf6 23.Rxh5+ Kg7 24.Rxe5 Kf7 25.g3 Rc6
26.Rb3 Rg8 27.Kg2 d4 28.c4 Rh6 29.Rb1 c5 30.Rbe1 Nc6 31.R5e4 Rf6 32.a3
Rg7 33.g4 Rh6 34.Rf4+ Kg8 35.Re8+ Kh7 36.Kg3 Ne7 37.h4 Re6 38.Re4 Rxe4
39.dxe4 Nc6 40.e5 Na5 41.e6 Nxc4 42.Rd8 Kg6 43.Rd7 Rg8
44.Rxa7 Kf6 45.e7 Re8 46.g5+ Kf5 47.a4 Ke6 48.f4 Rxe7 49.f5+ Kf7 50.Ra6
Re3+ 51.Kf4 Re2 52.Ra7+ Kf8 53.f6 Rxc2 54.g6 Rf2+ 55.Kg5 Rg2+ 56.Kh6
Rxg6+ 57.Kxg6 Ne5+ 58.Kf5 Nc6 59.Rc7 Nd8 60.h5 1-0
Rev-London Resigns here:
[fen]3n1k2/2R5/1p3P2/2p2K1P/P2p4/8/8/8 w - - 0 60[/fen]
Lets Reverse Colors next Regards
Steve
P.S.Full Pgn...
[White "London 68030"]
[Black "Revelation-London"]
[TimeControl "1 Min/Avg."]
[Result "1-0"]
1.e4 Nf6 2.Nc3 e5 3.Nf3 Nc6 4.Bb5 Bb4 5.O-O O-O 6.d3 Bxc3 7.bxc3 d6
8.Bg5 Bd7 9.Nd2 Ne7 10.Bxf6 gxf6 11.Bxd7 Qxd7 12.Qf3 Qe6 13.Rfb1 b6
14.Rb4 f5 15.exf5 Qxf5 16.Qxf5 Nxf5 17.Rg4+ Kh8 18.Ne4 f6 19.Rb1 h5
20.Rg6 d5 21.Nxf6 Ne7 22.Rg5 Rxf6 23.Rxh5+ Kg7 24.Rxe5 Kf7 25.g3 Rc6
26.Rb3 Rg8 27.Kg2 d4 28.c4 Rh6 29.Rb1 c5 30.Rbe1 Nc6 31.R5e4 Rf6 32.a3
Rg7 33.g4 Rh6 34.Rf4+ Kg8 35.Re8+ Kh7 36.Kg3 Ne7 37.h4 Re6 38.Re4 Rxe4
39.dxe4 Nc6 40.e5 Na5 41.e6 Nxc4 42.Rd8 Kg6 43.Rd7 Rg8
44.Rxa7 Kf6 45.e7 Re8 46.g5+ Kf5 47.a4 Ke6 48.f4 Rxe7 49.f5+ Kf7 50.Ra6
Re3+ 51.Kf4 Re2 52.Ra7+ Kf8 53.f6 Rxc2 54.g6 Rf2+ 55.Kg5 Rg2+ 56.Kh6
Rxg6+ 57.Kxg6 Ne5+ 58.Kf5 Nc6 59.Rc7 Nd8 60.h5 1-0