My EarlyWork.zip file

[Michael Sherwin]

I lost it due to my illness, carelessness, old floppy disk going bad and hard drive crashes. The zip file had one working winboard chess engine and three move generator examples with perft. Two of the move generator examples were very interesting. The one called Godzilla was 100% table driven assembler code linked with a C front end. And the other was an incremental update move generator. I am very curious how they would perform on modern processors. And also I might like to make them into minimal uci example engines. EarlyWork.zip was included with some versions of RomiChess. Does anyone still have that file that they can send to me? Thanks.

[Guenther]

I lost it due to my illness, carelessness, old floppy disk going bad and hard drive crashes. The zip file had one working winboard chess engine and three move generator examples with perft. Two of the move generator examples were very interesting. The one called Godzilla was 100% table driven assembler code linked with a C front end. And the other was an incremental update move generator. I am very curious how they would perform on modern processors. And also I might like to make them into minimal uci example engines. EarlyWork.zip was included with some versions of RomiChess. Does anyone still have that file that they can send to me? Thanks.

Found it! I will send it in a few minutes Michael.
Edit: will this old mail adress work: mjsherwin[at]netzero.net?

Cheers,
Guenther

[Michael Sherwin]

I lost it due to my illness, carelessness, old floppy disk going bad and hard drive crashes. The zip file had one working winboard chess engine and three move generator examples with perft. Two of the move generator examples were very interesting. The one called Godzilla was 100% table driven assembler code linked with a C front end. And the other was an incremental update move generator. I am very curious how they would perform on modern processors. And also I might like to make them into minimal uci example engines. EarlyWork.zip was included with some versions of RomiChess. Does anyone still have that file that they can send to me? Thanks.

Found it! I will send it in a few minutes Michael.

Cheers,
Guenther

Wow! Thanks Guenther, been a long time.

[velmarin]

Oh, yea!!
https://www.mediafire.com/?4npbipdezyjqhxm

[Michael Sherwin]

Oh, yea!!
https://www.mediafire.com/?4npbipdezyjqhxm

Hi Jose, thanks a million! To everyone!

[Guenther]

Oh, yea!!
https://www.mediafire.com/?4npbipdezyjqhxm

Hi Jose, thanks a million! To everyone!

Great you have it, because my mail bounced back.

Guenther

[Michael Sherwin]

I lost it due to my illness, carelessness, old floppy disk going bad and hard drive crashes. The zip file had one working winboard chess engine and three move generator examples with perft. Two of the move generator examples were very interesting. The one called Godzilla was 100% table driven assembler code linked with a C front end. And the other was an incremental update move generator. I am very curious how they would perform on modern processors. And also I might like to make them into minimal uci example engines. EarlyWork.zip was included with some versions of RomiChess. Does anyone still have that file that they can send to me? Thanks.

Got it and everything works.

On Athlon 3800 see readme.txt in zip
Conundrum > 16,000,000 nps
Godzilla > 20,000,000 nps

On i7 overclocked to 3.8 GHz
Conundrum > 39,000,000 nps
Godzilla > 67,000,000 nps

Seems like Intel has optimized this processor to do better at table driven assembly code!

[Michael Sherwin]

I called Conundrum an incremental update move generator. I'm not sure that terminology is correct. It does not do incremental move generation as understood by engine programmers. It creates a data structure of moves and updates that structure after every move by processing the from square and to squares and scanning for affected pieces. It is extremely complicated compared to normal move generators. It will take me awhile to wrap (or re wrap) my brain around it. And I'm not sure what to call it.

[Norbert Raimund Leisner]

It is possible that a WinBoard compatible edition of Carnivor can be made?

The including version can be executed in console mode, but crashes under Arena-interface with CECP-protocol 1 respectively 2. My OS is Windows 7 32 bit.

Best wishes,
Norbert

[Michael Sherwin]

It is possible that a WinBoard compatible edition of Carnivor can be made?

The including version can be executed in console mode, but crashes under Arena-interface with CECP-protocol 1 respectively 2. My OS is Windows 7 32 bit.

Best wishes,
Norbert

I fixed the crashing bug once before. I will see if I still can figure out what was causing it and fix it again. After all this time it could take awhile.