Author: Gerd Isenberg
Date: 12:23:30 12/12/05
Go up one level in this thread
On December 12, 2005 at 12:18:50, Chrilly Donninger wrote:
>>>
>>>.text:0040D156 call sub_40D1B0
>>>.text:0040D15B mov esi, eax
>>>.text:0040D15D lea eax, [esp+278h+var_238]
>>>.text:0040D161 push eax
>>>.text:0040D162 push edi
>>>.text:0040D163 neg esi
>>>.text:0040D165 call sub_40A780
>>>.text:0040D16A add esp, 18h
>>>.text:0040D16D cmp esi, [esp+268h+var_25C]
>>>.text:0040D171 jle short loc_40D183
>>>.text:0040D173 cmp esi, ebx
>>>.text:0040D175 mov [esp+268h+var_25C], esi
>>>.text:0040D179 jle short loc_40D183
>>>.text:0040D17B cmp esi, [ebp+arg_4]
>>>.text:0040D17E mov [ebp+arg_0], esi
>>>.text:0040D181 jge short loc_40D13A
>>>
>>>
>>>Chrilly
>>
>>How can you tell? And how on earth did you find it?
>>
>>anthony
>
>There is a nice tool, called IDA-Pro. (IDA=Interactive DissAssembler).
>IDA produces the code above. And then one needs some (assembler) experience to
>read this.
>E.g. [ebp+arg_0] means the first argument to the function. And if one follows
>somewhat the logic, one sees that this is alpha, [ebp+arg_4] is argument 1 or
>beta.
>In Pseudo-Code the code above reads as:
>val=-Quiescence(...);
>UndoMove
>if(val>bestscore) { bestscore=val; }
>if(val>alpha) { alpha=val; }
>if(val>beta) { return val; } // Actually goto return-sequence.
>
Hi Chrilly,
a little nickpicking on your C-code ;-)
I translate the conditional part that way:
if (val > bestscore)
{
bestscore = val;
if(val > alpha)
{
alpha = val;
if(val >= beta)
return val; // Actually goto return-sequence.
}
}
i personally prefere this one, to save the assigment to alpha most of the time
for nullwindow cuts:
if (val > bestscore)
{
bestscore = val;
if(val >= beta)
return val; // Actually goto return-sequence.
if(val > alpha)
alpha = val;
}
May be Vas may archive another 0.1 ELO point with this improvement.
Cheers,
Gerd
>All programms are open-source. Some opensource is only more difficult to read
>than other.
>
>Dissassembling is no efficient way to improve one owns code. It does not help,
>if one knows, that programm X has a nice trick. The trick must fit in one owns
>programm/architecture. Usually it does not. E.g the well documented Fruit
>History-Pruning did not work at all in Hydra. Dissassembling is just an old
>hackers fun.
>
>Chrilly
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.