Christmas Engine Quiz

[Christopher Conkie]

While you are eating your Christmas pudding with all those lovely ingredients, we thought you might like a little something to do......



There is one question in our quiz.

Here goes.

Q.1 Which engine is this? (Obviously it's a recipe of personal taste....)

Code: Select all

.rdata:0000000000484090 0000000E C info depth %d                                                                                                                                                                                                                                               
.rdata:00000000004840A0 0000002D C info depth %d nodes %I64d time %.0f nps %.0f                                                                                                                                                                                                               
.rdata:00000000004840D0 00000036 C info depth %d nodes %I64d score cp %d time %.0f pv %s                                                                                                                                                                                                       
.rdata:0000000000484110 00000041 C info lowerbound depth %d nodes %I64d score cp %d time %.0f pv %s                                                                                                                                                                                           
.rdata:0000000000484160 00000041 C info upperbound depth %d nodes %I64d score cp %d time %.0f pv %s                                                                                                                                                                                           
.rdata:00000000004841A8 00000038 C info depth %d nodes %I64d score mate %d time %.0f pv %s                                                                                                                                                                                                     
.rdata:00000000004841E0 00000043 C info lowerbound depth %d nodes %I64d score mate %d time %.0f pv %s                                                                                                                                                                                         
.rdata:0000000000484230 00000043 C info upperbound depth %d nodes %I64d score mate %d time %.0f pv %s                                                                                                                                                                                         
.rdata:0000000000484278 00000023 C info currmovenumber %d currmove %s                                                                                                                                                                                                                     
.rdata:00000000004842A0 00000024 C info nodes %I64d time %.0f nps %.0f

Prize

One of these......

http://www.cakebaker.co.uk/bake-italian-christmas-cake.html



Christopher

[Norman Schmidt]

looks kinda like fruitcake (with a twist...)

...of course you can start with the basic recipe, change it a little, and end up with many flavorful and sumptuous offerings. i do especially like that italian christmas cake (with the candied fruit)....absolutely delicious!

[Christopher Conkie]

looks kinda like fruitcake (with a twist...)

...of course you can start with the basic recipe, change it a little, and end up with many flavorful and sumptuous offerings. i do especially like that italian christmas cake (with the candied fruit)....absolutely delicious!

No two cakes are the same.....I like your twist analagy.....

Still sticking with Fruit?

Code: Select all

.rdata:0000000000483A60 00000022 C option name %s type %s default %s                                                                                                                                                                                                                           
.rdata:0000000000483A88 00000025 C option name %s type %s default %s %s

Christopher

[El Gringo]

looks kinda like fruitcake (with a twist...)

...of course you can start with the basic recipe, change it a little, and end up with many flavorful and sumptuous offerings. i do especially like that italian christmas cake (with the candied fruit)....absolutely delicious!

No two cakes are the same.....I like your twist analagy.....

Still sticking with Fruit?

Code: Select all

.rdata:0000000000483A60 00000022 C option name %s type %s default %s                                                                                                                                                                                                                           
.rdata:0000000000483A88 00000025 C option name %s type %s default %s %s

Christopher

Hi,

Fish and fruit are not a good receipt. Never tried rybka with bananas


Best
Johan

[Christopher Conkie]

looks kinda like fruitcake (with a twist...)

...of course you can start with the basic recipe, change it a little, and end up with many flavorful and sumptuous offerings. i do especially like that italian christmas cake (with the candied fruit)....absolutely delicious!

No two cakes are the same.....I like your twist analagy.....

Still sticking with Fruit?

Code: Select all

.rdata:0000000000483A60 00000022 C option name %s type %s default %s                                                                                                                                                                                                                           
.rdata:0000000000483A88 00000025 C option name %s type %s default %s %s

Christopher

Hi,

Fish and fruit are not a good receipt. Never tried rybka with bananas


Best
Johan

I think I know what you mean Johan.....

Code: Select all

   if (pv[0] == move && move_is_ok(pv[1])) {
      move_to_string(pv[1],ponder_string,256);
      send("bestmove %s ponder %s",move_string,ponder_string);
   } else {
      send("bestmove %s",move_string);
   }

Or maybe even.....

Code: Select all

  if (best_pv[0] == best_move && best_pv[1] != 0) {
    move_to_string(best_pv[1], ponder);
    fprintf(stdout,"bestmove %s ponder %s\n",string,ponder);
  }
  else fprintf(stdout,"bestmove %s\n",string);

Or put another way......

Code: Select all

.rdata:00000000004842F0 00000016 C bestmove %s ponder %s                                                                                                                                                                                                                                       
.rdata:0000000000484308 0000000C C bestmove %s

Obviously there are many ways to do this.....one mans......

Code: Select all

info currmovenumber %d currmove %s

Is anothers.......

Code: Select all

info currmove %s currmovenumber %d

So to summarise so far......we have fruit and fish on the menu. Is that correct?



Christopher

[Christopher Conkie]

Ok...one final clue.....

Code: Select all

Found %d CPUs\n

Tell me what you SEE.



Christopher

[Norman Schmidt]

looks kinda like fruitcake (with a twist...)

...of course you can start with the basic recipe, change it a little, and end up with many flavorful and sumptuous offerings. i do especially like that italian christmas cake (with the candied fruit)....absolutely delicious!

No two cakes are the same.....I like your twist analagy.....

Still sticking with Fruit?

Code: Select all

.rdata:0000000000483A60 00000022 C option name %s type %s default %s                                                                                                                                                                                                                           
.rdata:0000000000483A88 00000025 C option name %s type %s default %s %s

Christopher

No,
it's slighty different...

fruit defines 64-bit integers as S64_FORMAT:

#ifdef _MSC_VER
# define S64_FORMAT "%I64d"
# define U64_FORMAT "%016I64X"
#else
# define S64_FORMAT "%lld"
# define U64_FORMAT "%016llX"
#endif

here is the code when fruit prints a PV:
(as far as I know the only engine to use upper and lowerbounds here)
if (false) {
} else if (flags == SearchExact) {
send("info depth %d seldepth %d score cp %d time %.0f nodes " S64_FORMAT " pv %s",depth,max_depth,value,time*1000.0,node_nb,pv_string);
} else if (flags == SearchLower) {
send("info depth %d seldepth %d score cp %d lowerbound time %.0f nodes " S64_FORMAT " pv %s",depth,max_depth,value,time*1000.0,node_nb,pv_string);
} else if (flags == SearchUpper) {
send("info depth %d seldepth %d score cp %d upperbound time %.0f nodes " S64_FORMAT " pv %s",depth,max_depth,value,time*1000.0,node_nb,pv_string);

here is the data you presented:

.rdata:0000000000484110 00000041 C info lowerbound depth %d nodes %I64d score cp %d time %.0f pv %s
.rdata:0000000000484160 00000041 C info upperbound depth %d nodes %I64d score cp %d time %.0f pv %s

[Norman Schmidt]

looks kinda like fruitcake (with a twist...)

...of course you can start with the basic recipe, change it a little, and end up with many flavorful and sumptuous offerings. i do especially like that italian christmas cake (with the candied fruit)....absolutely delicious!

No two cakes are the same.....I like your twist analagy.....

Still sticking with Fruit?

Code: Select all

.rdata:0000000000483A60 00000022 C option name %s type %s default %s                                                                                                                                                                                                                           
.rdata:0000000000483A88 00000025 C option name %s type %s default %s %s

Christopher

No,
it's slighty different...

fruit provides a define for 64-bit (signed and unsigned) integers S64_FORMAT andd U64_FORMAT:

#ifdef _MSC_VER
# define S64_FORMAT "%I64d"
# define U64_FORMAT "%016I64X"
#else
# define S64_FORMAT "%lld"
# define U64_FORMAT "%016llX"
#endif

for ex, here is the info string code from fruit::
(the only source code I've seen that uses upper and lowerbounds, but there's probably others)
if (false) {
} else if (flags == SearchExact) {
send("info depth %d seldepth %d score cp %d time %.0f nodes " S64_FORMAT " pv %s",depth,max_depth,value,time*1000.0,node_nb,pv_string);
} else if (flags == SearchLower) {
send("info depth %d seldepth %d score cp %d lowerbound time %.0f nodes " S64_FORMAT " pv %s",depth,max_depth,value,time*1000.0,node_nb,pv_string);
} else if (flags == SearchUpper) {
send("info depth %d seldepth %d score cp %d upperbound time %.0f nodes " S64_FORMAT " pv %s",depth,max_depth,value,time*1000.0,node_nb,pv_string);

(without this define, you'd basically have (an exact) match)

here is the data you presented (from a different UCI info string):

.rdata:0000000000484110 00000041 C info lowerbound depth %d nodes %I64d score cp %d time %.0f pv %s
.rdata:0000000000484160 00000041 C info upperbound depth %d nodes %I64d score cp %d time %.0f pv %s

[Christopher Conkie]

fruit defines 64-bit integers as S64_FORMAT:

Code: Select all

# define S64_FORMAT "%I64d"

Code: Select all

.rdata:0000000000484110 00000041 C info lowerbound depth %d nodes %I64d score cp %d time %.0f pv %s

In disassembly you will not see S64_FORMAT but instead %I64d as that is part of the string by definition.

But there is a very valid point that you made Norman.

Code: Select all

as far as I know the only engine to use upper and lowerbounds here

Yes....you are right about that. Still remains this although.....

Code: Select all

Found %d CPUs\n

i.e.

Is Fruit SMP in open source?

So where is that from? (only one place btw...well two if you count what i'm showing you).

You have earned so far....half an italian cake.



Christopher

[Norman Schmidt]

fruit defines 64-bit integers as S64_FORMAT:

Code: Select all

# define S64_FORMAT "%I64d"

Code: Select all

.rdata:0000000000484110 00000041 C info lowerbound depth %d nodes %I64d score cp %d time %.0f pv %s

In disassembly you will not see S64_FORMAT but instead %I64d as that is part of the string by definition.

But there is a very valid point that you made Norman.

Code: Select all

as far as I know the only engine to use upper and lowerbounds here

Yes....you are right about that. Still remains this although.....

Code: Select all

Found %d CPUs\n

i.e.

Is Fruit SMP in open source?

So where is that from? (only one place btw...well two if you count what i'm showing you).

You have earned so far....half an italian cake.



Christopher

fruit was not SMP, but toga14beta5c (based on Fruit) is and yes of course it's open-source.

Found CPUs must be part of a function that queries windows for the # of physical processors...

hence a (configurable) top limit would be needed, something like UCI option max_threads...i know that max_threads exists in Grapefruit, Rybka, and there are probably others as well, ...?

[Christopher Conkie]

fruit defines 64-bit integers as S64_FORMAT:

Code: Select all

# define S64_FORMAT "%I64d"

Code: Select all

.rdata:0000000000484110 00000041 C info lowerbound depth %d nodes %I64d score cp %d time %.0f pv %s

In disassembly you will not see S64_FORMAT but instead %I64d as that is part of the string by definition.

But there is a very valid point that you made Norman.

Code: Select all

as far as I know the only engine to use upper and lowerbounds here

Yes....you are right about that. Still remains this although.....

Code: Select all

Found %d CPUs\n

i.e.

Is Fruit SMP in open source?

So where is that from? (only one place btw...well two if you count what i'm showing you).

You have earned so far....half an italian cake.



Christopher

fruit was not SMP, but toga14beta5c (based on Fruit) is and yes of course it's open-source.

Never tried Toga (its against my Christmas religion) but.....could i assume that it plays well on up to 8 CPUs?

Just a guess.....



Christopher

[Norman Schmidt]

fruit defines 64-bit integers as S64_FORMAT:

Code: Select all

# define S64_FORMAT "%I64d"

Code: Select all

.rdata:0000000000484110 00000041 C info lowerbound depth %d nodes %I64d score cp %d time %.0f pv %s

In disassembly you will not see S64_FORMAT but instead %I64d as that is part of the string by definition.

But there is a very valid point that you made Norman.

Code: Select all

as far as I know the only engine to use upper and lowerbounds here

Yes....you are right about that. Still remains this although.....

Code: Select all

Found %d CPUs\n

i.e.

Is Fruit SMP in open source?

So where is that from? (only one place btw...well two if you count what i'm showing you).

You have earned so far....half an italian cake.



Christopher

fruit was not SMP, but toga14beta5c (based on Fruit) is and yes of course it's open-source.

Never tried Toga (its against my Christmas religion) but.....could i assume that it plays well on up to 8 CPUs?

Just a guess.....



Christopher

it would play extremely well....


here is the function from grapefruit...

// init_threads()

void init_threads( bool call_option )
{
SYSTEM_INFO sys_info;
int thread_count, option_max;
GetSystemInfo(&sys_info);
thread_count = sys_info.dwNumberOfProcessors;

if (thread_count > 16)
thread_count = 16;

if (!MP)
thread_count = 1;

if (call_option)
{
option_max = option_get_int("Max threads");

if (thread_count > option_max)
thread_count = option_max;
}
number_threads = thread_count;
}

(note : it does not print out the # of CPUs found)
therefore i'm ruling out grapefruit

[Christopher Conkie]

extremely well....


here is the function from grapefruit...
(note : it does not print out the # of CPUs found)

therefore i'm ruling out grapefruit

Code: Select all

Found %d CPUs\n

Code: Select all

printf("Found %d CPUs\n", nproc);

If I wanted to make a clone that was SMP, x64, and Fruit, I would look two engines....

Toga and.........Sloppy. (OK so I might look at Viper/Glaurung also....if the old 8 cpu thing is a determining factor)

Do you think that would be a good hypothesis to make?



I remember long ago during the WCCC (couple of years or so) that we told everyone that we felt that Lion++ was Togaish.....the rest, as they say is.....history. I can specifically remember having conversations about this with David Levy and I remember SMK looked at the code. It was found to be 80% Fruity IIRC and was disqualified from WCCC.

Now this a couple of days ago......

http://talkchess.com/forum/viewtopic.php?t=24822

Meanwhile we got a "new" engine.....from Italy.

Your thoughts please...for example....do you want to know the programmers name? The rest of the cake is in it for you if you can guess.....



Christopher

[Norman Schmidt]

extremely well....


here is the function from grapefruit...
(note : it does not print out the # of CPUs found)

therefore i'm ruling out grapefruit

Code: Select all

Found %d CPUs\n

Code: Select all

printf("Found %d CPUs\n", nproc);

If I wanted to make a clone that was SMP, x64, and Fruit, I would look two engines....

Toga and.........Sloppy. (OK so I might look at Viper/Glaurung also....if the old 8 cpu thing is a determining factor)

Do you think that would be a good hypothesis to make?



I remember long ago during the WCCC (couple of years or so) that we told everyone that we felt that Lion++ was Togaish.....the rest, as they say is.....history. I can specifically remember having conversations about this with David Levy and I remember SMK looked at the code. It was found to be 80% Fruity IIRC and was disqualified from WCCC.

Now this a couple of days ago......

http://talkchess.com/forum/viewtopic.php?t=24822

Meanwhile we got a "new" engine.....from Italy.

Your thoughts please...for example....do you want to know the programmers name? The rest of the cake is in it for you if you can guess.....



Christopher

No I can't guess, i didn't even know there was a new Italian engine..

I suggest:
verbally ask all suspects to write "conditional weight"...
whoever writes "Conditionated Weihgt" is your man!

<:

[Norman Schmidt]

May i ask, what is the name of this 'new' engine to which you're referring?

i.e. aren't you going to give us the answer to the quiz?
<:

[Christopher Conkie]

May i ask, what is the name of this 'new' engine to which you're referring?

i.e. aren't you going to give us the answer to the quiz?
<:

The name of the engine is Cipollino.

But......

I have made you the winner on the basis that you called it Fruit and let's face it......if anyone knows......you would.

Christmas List

Norman Schmidt - One Italian Cake.



Christopher

[Olivier Deville]

May i ask, what is the name of this 'new' engine to which you're referring?

i.e. aren't you going to give us the answer to the quiz?
<:

Hi Norm

The engine is not released at the moment.

Chris is checking the exe with permission from the author.

Olivier

[Christopher Conkie]

May i ask, what is the name of this 'new' engine to which you're referring?

i.e. aren't you going to give us the answer to the quiz?
<:

Hi Norm

The engine is not released at the moment.

Chris is checking the exe with permission from the author.

Olivier

We have finished checking. (Norm and I)



Christopher

[Norman Schmidt]

May i ask, what is the name of this 'new' engine to which you're referring?

i.e. aren't you going to give us the answer to the quiz?
<:

The name of the engine is Cipollino.

But......

I have made you the winner on the basis that you called it Fruit and let's face it......if anyone knows......you would.

Christmas List

Norman Schmidt - One Italian Cake.



Christopher

much experience in this area... (fruit cake that is!) <:

As for the Italian cake...i'd like to donate this generous prize (or any proceeds that arise from an eventual auction or sale of the cake) to Olivier and Chesswar...!

[Norman Schmidt]

i'm a bit uneasy here...i have only seen what's been posted so far in this thread. i would hate to see any unintentional injustice done to Cipollino (or any other engine for that matter).

maybe others need to be involved ... i.e. reach a consensus, and carefully consider if this constitutes irrefutable proof, or is it only enough to be strongly suspicious, or what...?

i know there's no such thing as 100% proof, so i realize a less-than-perfect decision has to be made at one point...