ElChinito: copies NextMove() from Crafty

[Paul Hunter]

Geschrieben von:/Posted by: Paul Hunter at 22 August 2004 23:58:56:

I posted this on CCC in a reply to Bob. I just thought folks here might be interested too.
Below is the Crafty function NextMove() and disassembly of the corresponding ElChinito code. I have added a few comments to the code so people can easily recognize it. Adding more comments is very easy for anyone.
Again, notice that the ElChinito code is from the Crafty code.
There is another thing peculiar to Crafty here. It's not a bug. But if you are not careful in copying the Crafty code, then it becomes your bug. In the case of ElChinito, it again copied it without understanding what it is for. Hence it becomes an ElChinito bug. I will leave that puzzle for some other time to warn future cloners.
#define NONE 0
#define HASH_MOVE 1
#define GENERATE_CAPTURE_MOVES 2
#define CAPTURE_MOVES 3
#define KILLER_MOVE_1 4
#define KILLER_MOVE_2 5
#define GENERATE_ALL_MOVES 6
#define SORT_ALL_MOVES 7
#define HISTORY_MOVES_1 8
#define HISTORY_MOVES_2 9
#define REMAINING_MOVES 10
#define ROOT_MOVES 11
int NextMove(TREE * tree, int ply, int wtm)
{
register int *bestp, *movep, *sortv;
register int history_value, bestval, index;
switch (tree->next_status[ply].phase) {
case HASH_MOVE:
tree->next_status[ply].phase = GENERATE_CAPTURE_MOVES;
if (tree->hash_move[ply]) {
tree->current_move[ply] = tree->hash_move[ply];
if (ValidMove(tree, ply, wtm, tree->current_move[ply]))
return (HASH_MOVE);
else
Print(128, "bad move from hash table, ply=%d\n", ply);
}
case GENERATE_CAPTURE_MOVES:
tree->next_status[ply].phase = CAPTURE_MOVES;
tree->last[ply] = GenerateCaptures(tree, ply, wtm,
tree->last[ply - 1]);
tree->next_status[ply].remaining = 0;
if (tree->hash_move[ply]) {
for (movep = tree->last[ply - 1], sortv = tree->sort_value;
movep < tree->last[ply]; movep++, sortv++)
if (*movep == tree->hash_move[ply]) {
*sortv = -999999;
*movep = 0;
tree->hash_move[ply] = 0;
} else {
if (p_values[Piece(*movep) + 7] next_status[ply].remaining++;
} else {
*sortv = Swap(tree, From(*movep), To(*movep), wtm);
if (*sortv >= 0)
tree->next_status[ply].remaining++;
}
}
} else {
for (movep = tree->last[ply - 1], sortv = tree->sort_value;
movep < tree->last[ply]; movep++, sortv++)
if (p_values[Piece(*movep) + 7] next_status[ply].remaining++;
} else {
*sortv = Swap(tree, From(*movep), To(*movep), wtm);
if (*sortv >= 0)
tree->next_status[ply].remaining++;
}
}
if (tree->last[ply] > tree->last[ply - 1] + 1) {
int temp1,
temp2,
*tmovep,
*tsortv;
int *end;
sortv = tree->sort_value + 1;
movep = tree->last[ply - 1] + 1;
end = tree->last[ply];
for (; movep < end; movep++, sortv++) {
temp1 = *movep;
temp2 = *sortv;
tmovep = movep - 1;
tsortv = sortv - 1;
while (tmovep >= tree->last[ply - 1] && *tsortv < temp2) {
*(tsortv + 1) = *tsortv;
*(tmovep + 1) = *tmovep;
tmovep--;
tsortv--;
}
*(tmovep + 1) = temp1;
*(tsortv + 1) = temp2;
}
}
tree->next_status[ply].last = tree->last[ply - 1];
case CAPTURE_MOVES:
if (tree->next_status[ply].remaining) {
tree->current_move[ply] = *(tree->next_status[ply].last);
*tree->next_status[ply].last++ = 0;
tree->next_status[ply].remaining--;
if (!tree->next_status[ply].remaining)
tree->next_status[ply].phase = KILLER_MOVE_1;
return (CAPTURE_MOVES);
}
tree->next_status[ply].phase = KILLER_MOVE_1;
case KILLER_MOVE_1:
if ((tree->hash_move[ply] != tree->killers[ply].move1) &&
ValidMove(tree, ply, wtm, tree->killers[ply].move1)) {
tree->current_move[ply] = tree->killers[ply].move1;
tree->next_status[ply].phase = KILLER_MOVE_2;
return (KILLER_MOVE_1);
}
case KILLER_MOVE_2:
if ((tree->hash_move[ply] != tree->killers[ply].move2) &&
ValidMove(tree, ply, wtm, tree->killers[ply].move2)) {
tree->current_move[ply] = tree->killers[ply].move2;
tree->next_status[ply].phase = GENERATE_ALL_MOVES;
return (KILLER_MOVE_2);
}
tree->next_status[ply].phase = GENERATE_ALL_MOVES;
case GENERATE_ALL_MOVES:
tree->last[ply] = GenerateNonCaptures(tree, ply, wtm, tree->last[ply]);
tree->next_status[ply].phase = HISTORY_MOVES_1;
case HISTORY_MOVES_1:
tree->next_status[ply].remaining = 1;
tree->next_status[ply].phase = HISTORY_MOVES_2;
bestval = 0;
bestp = 0;
for (movep = tree->last[ply - 1]; movep < tree->last[ply]; movep++)
if (*movep) {
if (*movep == tree->hash_move[ply] ||
*movep == tree->killers[ply].move1 ||
*movep == tree->killers[ply].move2)
*movep = 0;
else {
index = *movep & 4095;
history_value = (wtm) ? history_w[index] :
history_b[index];
if (history_value > bestval) {
bestval = history_value;
bestp = movep;
}
}
}
if (bestp) {
tree->current_move[ply] = *bestp;
*bestp = 0;
return (HISTORY_MOVES_1);
}
goto remaining_moves;
case HISTORY_MOVES_2:
bestval = 0;
bestp = 0;
for (movep = tree->last[ply - 1]; movep < tree->last[ply]; movep++)
if (*movep) {
index = *movep & 4095;
history_value = (wtm) ? history_w[index] : history_b[index];
if (history_value > bestval) {
bestval = history_value;
bestp = movep;
}
}
if (bestval) {
tree->current_move[ply] = *bestp;
*bestp = 0;
tree->next_status[ply].remaining++;
if (tree->next_status[ply].remaining > 3) {
tree->next_status[ply].phase = REMAINING_MOVES;
tree->next_status[ply].last = tree->last[ply - 1];
}
return (HISTORY_MOVES_2);
}
remaining_moves:
tree->next_status[ply].phase = REMAINING_MOVES;
tree->next_status[ply].last = tree->last[ply - 1];
case REMAINING_MOVES:
for (; tree->next_status[ply].last < tree->last[ply];
tree->next_status[ply].last++)
if (*tree->next_status[ply].last) {
tree->current_move[ply] = *tree->next_status[ply].last;
*tree->next_status[ply].last++ = 0;
return (REMAINING_MOVES);
}
return (NONE);
default:
Print(4095, "oops! next_status.phase is bad! [normal %d]\n",
tree->next_status[ply].phase);
return (NONE);
}
}
// NextMove() from ElChinito 3.25
// -------------------------------
Fun42f9a0 :: ; proc near // NextMove()
sub esp , 0Ch
push ebx
push ebp
mov ebp , dword ptr [ esp + 24 ]
push esi
push edi
lea edi , dword ptr [ ebp + 2 * ebp ]
shl edi , 02h
mov eax , dword ptr [ edi + 04B35C8h ]
dec eax
cmp eax , 09h
mov dword ptr [ esp + 20 ] , edi
ja Label43009e
jmp dword ptr [ 4 * eax + 04300D8h ]
// case HASH_MOVE:
mov dword ptr [ edi + 04B35C8h ] , 02h
mov eax , dword ptr [ 4 * ebp + 04B47A8h ]
test eax , eax
je Label42fa1c
push eax
mov dword ptr [ 4 * ebp + 04B46A4h ] , eax
mov eax , dword ptr [ esp + 40 ]
push eax
push ebp
call Fun4308d0 // call ValidMove()
add esp , 0Ch
test eax , eax
je Label42fa09
pop edi
pop esi
pop ebp
mov eax , 01h // return (HASH_MOVE);
pop ebx
add esp , 0Ch
ret
Label42fa09 ::
push ebp
pushd 0440558h
pushd 080h
call Fun405030
add esp , 0Ch
// case GENERATE_CAPTURE_MOVES:
Label42fa1c ::
mov edx , dword ptr [ esp + 36 ]
mov dword ptr [ edi + 04B35C8h ] , 03h
mov ecx , dword ptr [ 4 * ebp + 04B48A8h ]
push ecx
push edx
push ebp
call Fun4144a0
mov dword ptr [ 4 * ebp + 04B48ACh ] , eax
mov dword ptr [ edi + 04B35CCh ] , 00h
mov eax , dword ptr [ 4 * ebp + 04B47A8h ]
add esp , 0Ch
test eax , eax
je Label42fc0d
mov ecx , dword ptr [ 4 * ebp + 04B48A8h ]
cmp ecx , dword ptr [ 4 * ebp + 04B48ACh ]
mov dword ptr [ esp + 16 ] , ecx
mov esi , 04BDEFCh
jnb Label42fd3c
Label42fa79 ::
mov eax , dword ptr [ecx]
cmp eax , dword ptr [ 4 * ebp + 04B47A8h ]
jne Label42faa0
mov dword ptr [esi] , 0FFF0BDC1h // *sortv = -999999;
mov dword ptr [ecx] , 00h // *movep = 0;
mov dword ptr [ 4 * ebp + 04B47A8h ] , 00h // tree->hash_move[ply] = 0;
jmp Label42fb50
Label42faa0 ::
mov edx , dword ptr [ esp + 36 ]
mov ecx , eax
sar ecx , 06h
and ecx , 03Fh
test edx , edx
je Label42fb72
and ecx , 0FFFFh
mov dl , byte ptr [ ecx + 04BF860h ]
test dl , dl
mov dword ptr [ esp + 24 ] , ecx
mov ecx , eax
je Label42fbe5
mov ebp , dword ptr [ esp + 24 ]
xor ebx , ebx
mov bl , byte ptr [ ebp + 04BFD00h ]
sar ecx , 0Ch
and ecx , 07h
movsx edi ,word ptr [ 2 * ecx + 043EA70h ]
and edx , 0FFh
sub ebx , edi
mov edi , dword ptr [ 4 * ecx + 043C7D4h ]
shl edx , 08h
add ebx , edx
mov edx , dword ptr [ 4 * ebx + 06741C0h ]
sub edx , edi
jns Label42fb24
movsx ecx ,word ptr [ 2 * ecx + 043EA80h ]
sar eax , 0Fh
and eax , 07h
mov eax , dword ptr [ 4 * eax + 043C7D4h ]
add eax , ecx
add eax , edx
mov dword ptr [esi] , eax
jmp Label42fb3d
Label42fb24 ::
movsx ecx ,word ptr [ 2 * ecx + 043EA80h ]
sar eax , 0Fh
and eax , 07h
mov edx , dword ptr [ 4 * eax + 043C7D4h ]
add edx , ecx
mov dword ptr [esi] , edx
Label42fb3d ::
cmp dword ptr [esi] , 00h
mov edi , dword ptr [ esp + 20 ]
jl Label42fb4c
inc dword ptr [ edi + 04B35CCh ]
Label42fb4c ::
mov ebp , dword ptr [ esp + 32 ]
Label42fb50 ::
mov ecx , dword ptr [ esp + 16 ]
mov eax , dword ptr [ 4 * ebp + 04B48ACh ]
add ecx , 04h
add esi , 04h
cmp ecx , eax
mov dword ptr [ esp + 16 ] , ecx
jb Label42fa79
jmp Label42fd3c
Label42fb72 ::
and ecx , 0FFFFh
mov dl , byte ptr [ ecx + 04BFD00h ]
test dl , dl
mov dword ptr [ esp + 24 ] , ecx
mov ecx , eax
je Label42fbe5
mov ebp , dword ptr [ esp + 24 ]
xor ebx , ebx
mov bl , byte ptr [ ebp + 04BF860h ]
sar ecx , 0Ch
and ecx , 07h
movsx edi ,word ptr [ 2 * ecx + 043EA70h ]
and edx , 0FFh
sub ebx , edi
mov edi , dword ptr [ 4 * ecx + 043C7D4h ]
shl edx , 08h
add ebx , edx
mov edx , dword ptr [ 4 * ebx + 06741C0h ]
sub edx , edi
jns Label42fb24
movsx ecx ,word ptr [ 2 * ecx + 043EA80h ]
sar eax , 0Fh
and eax , 07h
mov eax , dword ptr [ 4 * eax + 043C7D4h ]
add eax , ecx
add eax , edx
mov dword ptr [esi] , eax
jmp Label42fb3d
Label42fbe5 ::
sar ecx , 0Ch
and ecx , 07h
movsx edx ,word ptr [ 2 * ecx + 043EA80h ]
sar eax , 0Fh
and eax , 07h
add edx , dword ptr [ 4 * eax + 043C7D4h ]
mov dword ptr [esi] , edx
inc dword ptr [ edi + 04B35CCh ]
jmp Label42fb50
Label42fc0d ::
mov eax , dword ptr [ 4 * ebp + 04B48A8h ]
cmp eax , dword ptr [ 4 * ebp + 04B48ACh ]
mov ebx , 04BDEFCh
mov dword ptr [ esp + 16 ] , eax
mov dword ptr [ esp + 32 ] , ebx
jnb Label42fd3c
Label42fc2e ::
mov eax , dword ptr [ esp + 16 ]
mov eax , dword ptr [eax]
mov edx , dword ptr [ esp + 36 ]
mov ecx , eax
sar ecx , 06h
and ecx , 03Fh
test edx , edx
mov esi , ecx
je Label42fc6c
and esi , 0FFFFh
mov dl , byte ptr [ esi + 04BF860h ]
test dl , dl
mov ecx , eax
je Label42fcf8
sar ecx , 0Ch
and ecx , 07h
xor ebx , ebx
mov bl , byte ptr [ esi + 04BFD00h ]
jmp Label42fc8c
Label42fc6c ::
and esi , 0FFFFh
mov dl , byte ptr [ esi + 04BFD00h ]
test dl , dl
mov ecx , eax
je Label42fcf8
sar ecx , 0Ch
and ecx , 07h
xor ebx , ebx
mov bl , byte ptr [ esi + 04BF860h ]
Label42fc8c ::
movsx esi ,word ptr [ 2 * ecx + 043EA70h ]
and edx , 0FFh
sub ebx , esi
mov esi , dword ptr [ 4 * ecx + 043C7D4h ]
shl edx , 08h
add ebx , edx
mov edx , dword ptr [ 4 * ebx + 06741C0h ]
sub edx , esi
jns Label42fcd4
movsx ecx ,word ptr [ 2 * ecx + 043EA80h ]
mov ebx , dword ptr [ esp + 32 ]
sar eax , 0Fh
and eax , 07h
mov eax , dword ptr [ 4 * eax + 043C7D4h ]
add eax , ecx
add eax , edx
mov dword ptr [ebx] , eax
jmp Label42fcf1
Label42fcd4 ::
movsx ecx ,word ptr [ 2 * ecx + 043EA80h ]
mov ebx , dword ptr [ esp + 32 ]
sar eax , 0Fh
and eax , 07h
mov edx , dword ptr [ 4 * eax + 043C7D4h ]
add edx , ecx
mov dword ptr [ebx] , edx
Label42fcf1 ::
cmp dword ptr [ebx] , 00h
jl Label42fd1b
jmp Label42fd15
Label42fcf8 ::
sar ecx , 0Ch
and ecx , 07h
movsx edx ,word ptr [ 2 * ecx + 043EA80h ]
sar eax , 0Fh
and eax , 07h
add edx , dword ptr [ 4 * eax + 043C7D4h ]
mov dword ptr [ebx] , edx
Label42fd15 ::
inc dword ptr [ edi + 04B35CCh ]
Label42fd1b ::
mov eax , dword ptr [ esp + 16 ]
mov ecx , dword ptr [ 4 * ebp + 04B48ACh ]
add eax , 04h
add ebx , 04h
cmp eax , ecx
mov dword ptr [ esp + 16 ] , eax
mov dword ptr [ esp + 32 ] , ebx
jb Label42fc2e
Label42fd3c ::
mov eax , dword ptr [ 4 * ebp + 04B48A8h ]
lea edx , dword ptr [ eax + 4 ]
mov eax , dword ptr [ 4 * ebp + 04B48ACh ]
cmp edx , eax
mov dword ptr [ esp + 16 ] , eax
jnb Label42fdc3
lea eax , dword ptr [ edx - 4 ]
mov esi , 04BDF00h
mov dword ptr [ esp + 24 ] , eax
Label42fd61 ::
mov ecx , dword ptr [edx]
cmp eax , dword ptr [ 4 * ebp + 04B48A8h ]
mov dword ptr [ esp + 20 ] , ecx
mov ecx , dword ptr [esi]
mov dword ptr [ esp + 32 ] , ecx
lea ecx , dword ptr [ esi - 4 ]
jb Label42fd9c
Label42fd79 ::
mov ebx , dword ptr [ esp + 32 ]
cmp dword ptr [ecx] , ebx
jnl Label42fd9c
mov ebx , dword ptr [ecx]
mov dword ptr [ ecx + 4 ] , ebx
mov ebx , dword ptr [eax]
mov dword ptr [ eax + 4 ] , ebx
mov ebx , dword ptr [ 4 * ebp + 04B48A8h ]
sub eax , 04h
sub ecx , 04h
cmp eax , ebx
jnb Label42fd79
Label42fd9c ::
mov ebx , dword ptr [ esp + 20 ]
mov dword ptr [ eax + 4 ] , ebx
mov eax , dword ptr [ esp + 32 ]
mov dword ptr [ ecx + 4 ] , eax
mov eax , dword ptr [ esp + 24 ]
mov ecx , dword ptr [ esp + 16 ]
add edx , 04h
add eax , 04h
add esi , 04h
cmp edx , ecx
mov dword ptr [ esp + 24 ] , eax
jb Label42fd61
Label42fdc3 ::
mov ecx , dword ptr [ 4 * ebp + 04B48A8h ]
mov dword ptr [ edi + 04B35D0h ] , ecx
mov eax , dword ptr [ edi + 04B35CCh ]
test eax , eax
je Label42fde3
mov byte ptr [ ebp + 04962C0h ] , 01h
jmp Label42fdea
Label42fde3 ::
mov byte ptr [ ebp + 04962C0h ] , 00h
// case CAPTURE_MOVES:
Label42fdea ::
mov eax , dword ptr [ edi + 04B35CCh ]
test eax , eax
je Label42fe3d
mov edx , dword ptr [ edi + 04B35D0h ]
mov eax , dword ptr [edx]
mov dword ptr [ 4 * ebp + 04B46A4h ] , eax
mov ecx , dword ptr [ edi + 04B35D0h ]
mov dword ptr [ecx] , 00h
mov edx , dword ptr [ edi + 04B35D0h ]
mov eax , 04h
add edx , eax
mov dword ptr [ edi + 04B35D0h ] , edx
dec dword ptr [ edi + 04B35CCh ]
jne Label42fe30
mov dword ptr [ edi + 04B35C8h ] , eax
Label42fe30 ::
pop edi
pop esi
pop ebp
mov eax , 03h // return (CAPTURE_MOVES);
pop ebx
add esp , 0Ch
ret
Label42fe3d ::
mov dword ptr [ edi + 04B35C8h ] , 04h
// case KILLER_MOVE_1:
mov eax , dword ptr [ 8 * ebp + 04B8CF4h ]
cmp dword ptr [ 4 * ebp + 04B47A8h ] , eax
je Label42fe8f
mov ebx , dword ptr [ esp + 36 ]
push eax
push ebx
push ebp
call Fun4308d0
add esp , 0Ch
test eax , eax
je Label42fe93
mov edx , dword ptr [ 8 * ebp + 04B8CF4h ]
mov dword ptr [ 4 * ebp + 04B46A4h ] , edx
mov dword ptr [ edi + 04B35C8h ] , 05h // tree->next_status[ply].phase =
KILLER_MOVE_2;
pop edi
pop esi
pop ebp
mov eax , 04h // return (KILLER_MOVE_1);
pop ebx
add esp , 0Ch
ret
Label42fe8f ::
mov ebx , dword ptr [ esp + 36 ]
// case KILLER_MOVE_2:
Label42fe93 ::
mov eax , dword ptr [ 8 * ebp + 04B8CF8h ]
cmp dword ptr [ 4 * ebp + 04B47A8h ] , eax
je Label42fed7
push eax
push ebx
push ebp
call Fun4308d0
add esp , 0Ch
test eax , eax
je Label42fed7
mov eax , dword ptr [ 8 * ebp + 04B8CF8h ]
mov dword ptr [ 4 * ebp + 04B46A4h ] , eax
mov dword ptr [ edi + 04B35C8h ] , 06h
pop edi
pop esi
pop ebp
mov eax , 05h
pop ebx
add esp , 0Ch
ret
Label42fed7 ::
mov dword ptr [ edi + 04B35C8h ] , 06h
jmp Label42fee7
mov ebx , dword ptr [ esp + 36 ]
Label42fee7 ::
// case GENERATE_ALL_MOVES:
mov ecx , dword ptr [ 4 * ebp + 04B48ACh ]
push ecx
push ebx
push ebp
call Fun416260
mov dword ptr [ 4 * ebp + 04B48ACh ] , eax
add esp , 0Ch
mov dword ptr [ edi + 04B35C8h ] , 08h
jmp Label42ff10
mov ebx , dword ptr [ esp + 36 ]
Label42ff10 ::
mov dword ptr [ edi + 04B35CCh ] , 01h
lea eax , dword ptr [ 4 * ebp + 04B48A8h ]
mov dword ptr [ edi + 04B35C8h ] , 09h
mov ecx , dword ptr [eax]
mov dword ptr [ esp + 32 ] , eax
mov eax , dword ptr [ 4 * ebp + 04B48ACh ]
xor esi , esi
xor edx , edx
cmp ecx , eax
jnb Label43005c
Label42ff44 ::
mov eax , dword ptr [ecx]
test eax , eax
je Label42ff8e
cmp eax , dword ptr [ 4 * ebp + 04B47A8h ]
je Label42ff88
cmp eax , dword ptr [ 8 * ebp + 04B8CF4h ]
je Label42ff88
cmp eax , dword ptr [ 8 * ebp + 04B8CF8h ]
je Label42ff88
and eax , 0FFFh
test ebx , ebx
je Label42ff77
mov eax , dword ptr [ 4 * eax + 06CD1E0h ]
jmp Label42ff7e
Label42ff77 ::
mov eax , dword ptr [ 4 * eax + 06C4760h ]
Label42ff7e ::
cmp eax , esi
jle Label42ff8e
mov esi , eax
mov edx , ecx
jmp Label42ff8e
Label42ff88 ::
mov dword ptr [ecx] , 00h
Label42ff8e ::
mov eax , dword ptr [ 4 * ebp + 04B48ACh ]
add ecx , 04h
cmp ecx , eax
jb Label42ff44
test edx , edx
je Label43005c
mov eax , dword ptr [edx]
pop edi
mov dword ptr [ 4 * ebp + 04B46A4h ] , eax
pop esi
pop ebp
mov dword ptr [edx] , 00h
mov eax , 08h
pop ebx
add esp , 0Ch
ret
mov ecx , dword ptr [ 4 * ebp + 04B48A8h ]
mov ebx , dword ptr [ 4 * ebp + 04B48ACh ]
lea eax , dword ptr [ 4 * ebp + 04B48A8h ]
xor esi , esi
xor edx , edx
cmp ecx , ebx
mov dword ptr [ esp + 32 ] , eax
jnb Label43005c
Label42ffe1 ::
mov eax , dword ptr [ecx]
test eax , eax
je Label43000b
and eax , 0FFFh
cmp dword ptr [ esp + 36 ] , 00h
je Label42fffc
mov eax , dword ptr [ 4 * eax + 06CD1E0h ]
jmp Label430003
Label42fffc ::
mov eax , dword ptr [ 4 * eax + 06C4760h ]
Label430003 ::
cmp eax , esi
jle Label43000b
mov esi , eax
mov edx , ecx
Label43000b ::
add ecx , 04h
cmp ecx , ebx
jb Label42ffe1
test esi , esi
je Label43005c
mov ecx , dword ptr [edx]
mov dword ptr [ 4 * ebp + 04B46A4h ] , ecx
mov dword ptr [edx] , 00h
mov ecx , dword ptr [ edi + 04B35CCh ]
inc ecx
mov eax , ecx
cmp eax , 03h
mov dword ptr [ edi + 04B35CCh ] , ecx
jle Label43004f
mov edx , dword ptr [ esp + 32 ]
mov dword ptr [ edi + 04B35C8h ] , 0Ah // tree->next_status[ply].phase =
REMAINING_MOVES;
mov eax , dword ptr [edx]
mov dword ptr [ edi + 04B35D0h ] , eax
Label43004f ::
pop edi
pop esi
pop ebp
mov eax , 09h // return (HISTORY_MOVES_2);
pop ebx
add esp , 0Ch
ret
Label43005c ::
// (label) remaining_moves:
mov ecx , dword ptr [ esp + 32 ]
mov dword ptr [ edi + 04B35C8h ] , 0Ah
mov edx , dword ptr [ecx]
mov dword ptr [ edi + 04B35D0h ] , edx
mov eax , dword ptr [ edi + 04B35D0h ]
cmp eax , dword ptr [ 4 * ebp + 04B48ACh ]
jnb Label43009e
Label430081 ::
mov eax , dword ptr [ edi + 04B35D0h ]
cmp dword ptr [eax] , 00h
jne Label4300a8
add eax , 04h
mov dword ptr [ edi + 04B35D0h ] , eax
cmp eax , dword ptr [ 4 * ebp + 04B48ACh ]
jb Label430081
Label43009e ::
pop edi
pop esi
pop ebp
xor eax , eax
pop ebx
add esp , 0Ch
ret
Label4300a8 ::
mov ecx , dword ptr [ edi + 04B35D0h ]
mov edx , dword ptr [ecx]
mov dword ptr [ 4 * ebp + 04B46A4h ] , edx
mov eax , dword ptr [ edi + 04B35D0h ]
mov dword ptr [eax] , 00h
add dword ptr [ edi + 04B35D0h ] , 04h
pop edi
pop esi
pop ebp
mov eax , 0Ah
pop ebx
add esp , 0Ch
ret

[Matthias Gemuh]

Geschrieben von:/Posted by: Matthias Gemuh at 23 August 2004 00:39:12:
Als Antwort auf:/In reply to: ElChinito: copies NextMove() from Crafty geschrieben von:/posted by: Paul Hunter at 22 August 2004 23:58:56:



Hi Paul,
you have done a superb job in proving that ElChinito is a Crafty ... etc.
Anybody who is still waiting for more proofs and e-mails simply hate evidence.
Thanks,
Matthias.




BigLion + Taktix

[Uri Blass]

Geschrieben von:/Posted by: Uri Blass at 23 August 2004 01:14:02:
Als Antwort auf:/In reply to: Do some people hate evidence ? geschrieben von:/posted by: Matthias Gemuh at 23 August 2004 00:39:12:

Hi Paul,
you have done a superb job in proving that ElChinito is a Crafty ... etc.
Anybody who is still waiting for more proofs and e-mails simply hate evidence.
Thanks,
Matthias.

Yes
people hate evidence
I also hate evidence.
I know that movei is weaker than Crafty but I hoped that it is at least better than a buggy clone of Crafty.
I hate to see evidence that it is even weaker than a buggy clone of Crafty.
Uri

[Roger Brown]

Geschrieben von:/Posted by: Roger Brown at 23 August 2004 02:17:16:
Als Antwort auf:/In reply to: Re: Do some people hate evidence ? geschrieben von:/posted by: Uri Blass at 23 August 2004 01:14:02:

Yes
people hate evidence
I also hate evidence.
I know that movei is weaker than Crafty but I hoped that it is at least better than a buggy clone of Crafty.
I hate to see evidence that it is even weaker than a buggy clone of Crafty.
Uri

Awwwwwwww.....
I hope you recover from this let-down.

Incidentally Uri, could you send me an e-mail with a readme telling me how to set the parameters for the latest Movei beta you sent me?
Should you have a later model could you send that as well?
Thanks.
Later.

[Lance Perkins]

Geschrieben von:/Posted by: Lance Perkins at 23 August 2004 04:45:27:
Als Antwort auf:/In reply to: Do some people hate evidence ? geschrieben von:/posted by: Matthias Gemuh at 23 August 2004 00:39:12:

Yup. So don't present a DNA evidence. You will just be labeled as witch hunting.

Hi Paul,
you have done a superb job in proving that ElChinito is a Crafty ... etc.
Anybody who is still waiting for more proofs and e-mails simply hate evidence.
Thanks,
Matthias.

[Matthias Gemuh]

Geschrieben von:/Posted by: Matthias Gemuh at 23 August 2004 10:18:20:
Als Antwort auf:/In reply to: Re: Do some people hate evidence ? geschrieben von:/posted by: Uri Blass at 23 August 2004 01:14:02:

I know that movei is weaker than Crafty but I hoped that it is at least better than a buggy clone of Crafty.
I hate to see evidence that it is even weaker than a buggy clone of Crafty.
Uri

If you want to see evidence that current Movei is weaker than buggy clones,
simply stop the miraculous improvements in Movei .
Best,
Matthias.



BigLion + Taktix

[Uri Blass]

Geschrieben von:/Posted by: Uri Blass at 23 August 2004 13:06:32:
Als Antwort auf:/In reply to: Re: Do some people hate evidence ? geschrieben von:/posted by: Matthias Gemuh at 23 August 2004 10:18:20:

I know that movei is weaker than Crafty but I hoped that it is at least better than a buggy clone of Crafty.
I hate to see evidence that it is even weaker than a buggy clone of Crafty.
Uri

If you want to see evidence that current Movei is weaker than buggy clones,
simply stop the miraculous improvements in Movei .
Best,
Matthias.

I really stopped it in the last weeks and in the time that I worked on movei in the last weeks I only did cosmetic changes
to do the code more easy to read and better organized and some change in the time management code that probably not change much about playing strength.
There is still a lot of work to do in it and I also plan to do other things that are not going to improve significantly the strength of the engine in the near future.
I plan to work many hours on things like reading pgn files because I want to have the option to build a book for movei based on games and I do not like to use some external program to do things like translating games to full notation because it make the task of generating a book for movei more complicated.
I also plan to get rid of some stupid things that movei does that are not very important for games.
For example if movei finish to search during pondering because of mate or maximal depth or a single move it does not remember it's analysis and undo the move and wait passively for winboard commands.
It is not very important when movei is in book or have a single move to play but it is important when movei finds a mate because it means that movei may not respond immediatly to the expected move in case of seeing mate during pondering and I do not like it.
This is one change except enabling resign option that I plan to do in movei.
playing faster in some simple drawn endgames is also planned.
Uri

[Volker Boehm]

Geschrieben von:/Posted by: Volker Boehm at 23 August 2004 15:15:07:
Als Antwort auf:/In reply to: Re: Do some people hate evidence ? geschrieben von:/posted by: Uri Blass at 23 August 2004 13:06:32:

I know that movei is weaker than Crafty but I hoped that it is at least better than a buggy clone of Crafty.
I hate to see evidence that it is even weaker than a buggy clone of Crafty.
Uri

If you want to see evidence that current Movei is weaker than buggy clones,
simply stop the miraculous improvements in Movei .
Best,
Matthias.

I really stopped it in the last weeks and in the time that I worked on movei in the last weeks I only did cosmetic changes
to do the code more easy to read and better organized and some change in the time management code that probably not change much about playing strength.
There is still a lot of work to do in it and I also plan to do other things that are not going to improve significantly the strength of the engine in the near future.
I plan to work many hours on things like reading pgn files because I want to have the option to build a book for movei based on games and I do not like to use some external program to do things like translating games to full notation because it make the task of generating a book for movei more complicated.
I also plan to get rid of some stupid things that movei does that are not very important for games.
For example if movei finish to search during pondering because of mate or maximal depth or a single move it does not remember it's analysis and undo the move and wait passively for winboard commands.
It is not very important when movei is in book or have a single move to play but it is important when movei finds a mate because it means that movei may not respond immediatly to the expected move in case of seeing mate during pondering and I do not like it.
This is one change except enabling resign option that I plan to do in movei.
playing faster in some simple drawn endgames is also planned.
Uri

If you have a good implementation for the latter, just tell me. I have a loop at the end of the calculation thread. It works but it is ugly. I should use semaphores but I hadn´t the time to read how to implement:
// stop is set to true by the "reading" thread
while (!stop && pondering)
{
Something_Useless_To_Prevent_Optimizer_To_Scip_This_Code();
}

Greetings Volker

[Michael Scheidl]

Geschrieben von:/Posted by: Michael Scheidl at 23 August 2004 15:55:31:
Als Antwort auf:/In reply to: ElChinito: copies NextMove() from Crafty geschrieben von:/posted by: Paul Hunter at 22 August 2004 23:58:56:

Just wanted to ask if you did check other engines the same way but did NOT find any evidence or clone suspicion?
I'm not sure if you'd want to mention those investigated, which did not seem to contain crafty code (if you have such results yet), or if it would be fair to mention those, but OTOH it may help to end clone rumours for engines which are repeadetly accused...
Thanks anyway,
Mike Scheidl

[Uri Blass]

Geschrieben von:/Posted by: Uri Blass at 23 August 2004 16:30:19:
Als Antwort auf:/In reply to: Re: Do some people hate evidence ? geschrieben von:/posted by: Volker Boehm at 23 August 2004 15:15:07:

I know that movei is weaker than Crafty but I hoped that it is at least better than a buggy clone of Crafty.
I hate to see evidence that it is even weaker than a buggy clone of Crafty.
Uri

If you want to see evidence that current Movei is weaker than buggy clones,
simply stop the miraculous improvements in Movei .
Best,
Matthias.

I really stopped it in the last weeks and in the time that I worked on movei in the last weeks I only did cosmetic changes
to do the code more easy to read and better organized and some change in the time management code that probably not change much about playing strength.
There is still a lot of work to do in it and I also plan to do other things that are not going to improve significantly the strength of the engine in the near future.
I plan to work many hours on things like reading pgn files because I want to have the option to build a book for movei based on games and I do not like to use some external program to do things like translating games to full notation because it make the task of generating a book for movei more complicated.
I also plan to get rid of some stupid things that movei does that are not very important for games.
For example if movei finish to search during pondering because of mate or maximal depth or a single move it does not remember it's analysis and undo the move and wait passively for winboard commands.
It is not very important when movei is in book or have a single move to play but it is important when movei finds a mate because it means that movei may not respond immediatly to the expected move in case of seeing mate during pondering and I do not like it.
This is one change except enabling resign option that I plan to do in movei.
playing faster in some simple drawn endgames is also planned.
Uri

If you have a good implementation for the latter, just tell me. I have a loop at the end of the calculation thread. It works but it is ugly. I should use semaphores but I hadn´t the time to read how to implement:
// stop is set to true by the "reading" thread
while (!stop && pondering)
{
Something_Useless_To_Prevent_Optimizer_To_Scip_This_Code();
}

Greetings Volker

I do not understand what is semaphores.
I thought to wait passively for winboard command when I finish to ponder because of checkmate or book position or maximal depth and I do not see what is ugly in it.
I am afraid that I have no good implementation for nothing.
I have no threads in my code and I simply check every 8192 nodes for winboard commands.
Uri

[Roger Brown]

Geschrieben von:/Posted by: Roger Brown at 23 August 2004 17:51:02:
Als Antwort auf:/In reply to: Re: Did you check other engines? geschrieben von:/posted by: Michael Scheidl at 23 August 2004 15:55:31:

Just wanted to ask if you did check other engines the same way but did NOT find any evidence or clone suspicion?

Hello Michael,
I believe that the burden should not be on Paul to investigate but on the cloners to cease and desist...
Later.

[Michael Scheidl]

Geschrieben von:/Posted by: Michael Scheidl at 23 August 2004 20:36:32:
Als Antwort auf:/In reply to: Re: Did you check other engines? geschrieben von:/posted by: Roger Brown at 23 August 2004 17:51:02:

I believe that the burden should not be on Paul to investigate but on the cloners to cease and desist...

No burden, but I understand that he already has checked a couple of engines voluntarily, so I was curious which ones did not raise clone suspicion.
I don't actually demand the info, I just think it might be of general interest. Also, meanwhile I saw he has checked List 4.61 and found nothing.
Thanks anyway.
The ICGA should hire somebody like Paul, who is capable of providing meaningful facts without the need to acquire sources.
Ok, that's about anything I wanted to say about that matter...
Regards,
Mike

[Leo Dijksman]

Geschrieben von:/Posted by: Leo Dijksman at 23 August 2004 20:56:18:
Als Antwort auf:/In reply to: Re: Did you check other engines? geschrieben von:/posted by: Michael Scheidl at 23 August 2004 20:36:32:

I believe that the burden should not be on Paul to investigate but on the cloners to cease and desist...

No burden, but I understand that he already has checked a couple of engines voluntarily, so I was curious which ones did not raise clone suspicion.
I don't actually demand the info, I just think it might be of general interest. Also, meanwhile I saw he has checked List 4.61 and found nothing.
Thanks anyway.
The ICGA should hire somebody like Paul, who is capable of providing meaningful facts without the need to acquire sources.
Ok, that's about anything I wanted to say about that matter...
Regards,
Mike

Then they must get an exe of the playing engine, and the question is if should they have got it?
Best wishes,
Leo.



WBEC Ridderkerk homepage.

[Thomas Mayer]

Geschrieben von:/Posted by: Thomas Mayer at 23 August 2004 21:57:06:
Als Antwort auf:/In reply to: Re: Did you check other engines? geschrieben von:/posted by: Michael Scheidl at 23 August 2004 20:36:32:

Hi Michael,

I don't actually demand the info, I just think it might be of general
interest. Also, meanwhile I saw he has checked List 4.61 and found nothing.
Thanks anyway.
The ICGA should hire somebody like Paul, who is capable of providing
meaningful facts without the need to acquire sources.
Ok, that's about anything I wanted to say about that matter...

Well, when you find nothing the way Paul Hunter checks the code that does not mean that there IS nothing, it's no proof... But when he finds something then it is clear that there is something. Just to clearify that...
Greets, Thomas