Simple make question

[outAtime]

This should be an easy one for you experts out there. Im trying to make the second move of the PV, but I also need to make it somewhere other than in search.c.

This is what the code for update pv looks like in search():

Code: Select all

      /* update the pv: */
      pv[ply][ply] = moves[i];
      for &#40;j = ply+1; j < pv_length&#91;ply+1&#93;; j++)
      pv&#91;ply&#93;&#91;j&#93; = pv&#91;ply+1&#93;&#91;j&#93;;
      pv_length&#91;ply&#93; = pv_length&#91;ply+1&#93;;

I would try make (&moves[1], 0), but I dont think that would be right, and moves[MOVE_BUFF] isn't a global variable anyways. Maybe I need to make a new global variable? pv[PV_BUFF][PV_BUFF] is global, but Im not sure if I can use that one to make the second move in the PV.

The regular make in search is:

Code: Select all

 make (&moves[0], i);

Im trying to implement some pseudo-code suggested by H.G.Muller:

Code: Select all

// WB command loop:
while(1) {
  if(stm == computer) {
    Search();
    abort = 0;
    bestMove = PV[0]; **in my case "bestMove"is: comp_move = think();
    ponderMove = PV[1]; **haven't found a way other than flip the board
    MakeMove(bestMove); ** in my case: make (&comp_move, 0)
if I change the 0 to 1, it flips the board. 
    PrintMove(bestMove);
  }
  if(ponderOn) {
    if(ponderMove) {
      MakeMove(ponderMove);**Ive tried making new variables, none work
so far. They all just cause the board to flip 
and the engine has no idea about the first  PV move. 
      Search();
      UnMake();
    } else Search();
  }
  abort = 0;
  ReadLine(inbuf);
  // execute command
  ....
} 

Please help!! The above code is so simple, but I can't figure out how to make the second move in the PV!!!!!!!! For me PV[1], even if I make a global variable (I tried making a global variable pvmove[PV_BUFF]), just insn't working...do I need to do something in search() like:

Code: Select all

 moves[i] = pvmove[i] (or maybe pvmove[ply]???)

How do I make the second move of the PV?? In the pseudo-code example, how is PV[1] able to be used?

In my code: (not working! this is just an example of a recent experiment)

Code: Select all

  if ((allow_pondering && !force_mode && !result && game_ply != 0) 
        || is_analyzing)
	{
	***  ponder_move = pv[0][0]; ***
	  is_pondering = TRUE;
	  make (&ponder_move, 0);	    
	  think();
	  unmake (&ponder_move, 0); 	  
	  is_pondering = FALSE;

What should I change = pv[0][0] to, in order to have make (&ponder_move, 0) make the second move of the PV?

[bob]

This should be an easy one for you experts out there. Im trying to make the second move of the PV, but I also need to make it somewhere other than in search.c.

This is what the code for update pv looks like in search():

Code: Select all

      /* update the pv: */
      pv[ply][ply] = moves[i];
      for &#40;j = ply+1; j < pv_length&#91;ply+1&#93;; j++)
      pv&#91;ply&#93;&#91;j&#93; = pv&#91;ply+1&#93;&#91;j&#93;;
      pv_length&#91;ply&#93; = pv_length&#91;ply+1&#93;;

I would try make (&moves[1], 0), but I dont think that would be right, and moves[MOVE_BUFF] isn't a global variable anyways. Maybe I need to make a new global variable? pv[PV_BUFF][PV_BUFF] is global, but Im not sure if I can use that one to make the second move in the PV.

The regular make in search is:

Code: Select all

 make (&moves[0], i);

Im trying to implement some pseudo-code suggested by H.G.Muller:

Code: Select all

// WB command loop:
while(1) {
  if(stm == computer) {
    Search();
    abort = 0;
    bestMove = PV[0]; **in my case "bestMove"is: comp_move = think();
    ponderMove = PV[1]; **haven't found a way other than flip the board
    MakeMove(bestMove); ** in my case: make (&comp_move, 0)
if I change the 0 to 1, it flips the board. 
    PrintMove(bestMove);
  }
  if(ponderOn) {
    if(ponderMove) {
      MakeMove(ponderMove);**Ive tried making new variables, none work
so far. They all just cause the board to flip 
and the engine has no idea about the first  PV move. 
      Search();
      UnMake();
    } else Search();
  }
  abort = 0;
  ReadLine(inbuf);
  // execute command
  ....
} 

Please help!! The above code is so simple, but I can't figure out how to make the second move in the PV!!!!!!!! For me PV[1], even if I make a global variable (I tried making a global variable pvmove[PV_BUFF]), just insn't working...do I need to do something in search() like:

Code: Select all

 moves[i] = pvmove[i] (or maybe pvmove[ply]???)

How do I make the second move of the PV?? In the pseudo-code example, how is PV[1] able to be used?

In my code: (not working! this is just an example of a recent experiment)

Code: Select all

  if ((allow_pondering && !force_mode && !result && game_ply != 0) 
        || is_analyzing)
	{
	***  ponder_move = pv[0][0]; ***
	  is_pondering = TRUE;
	  make (&ponder_move, 0);	    
	  think();
	  unmake (&ponder_move, 0); 	  
	  is_pondering = FALSE;

What should I change = pv[0][0] to, in order to have make (&ponder_move, 0) make the second move of the PV?

If I am looking at your code correctly, pv[0] is the first move for any particular ply depth. pv[1] is the second, and so forth. So it looks to me like you want to use pv[something][1] to get to the second move. I think in the ponder case above, you want pv[0][1] it would seem. pv[0][0] should be the move you just played...

[outAtime]

Thankyou very much for the reply. I tried this out:

Code: Select all

          ponder_move = pv[0][1];
	  is_pondering = TRUE;
	  make (&ponder_move, 0);  
	  think();
	  unmake (&ponder_move, 0); 
	  is_pondering = FALSE;	

It compiles and the engine ponders, but (I was playing white pieces) as soon as the engine captured one of my pieces (a knight) the score jumped to +6.30 and it looked like it was thinking about taking my unprotected bishop next, which I think means it was thinking as though it was its turn to move, without being able to realize I could recapture with a pawn before it could take my bishop. Is that the way its supposed to work?
I think thats wrong, isn't it supposed to know the capturing piece can be recaptured during ponder?

[outAtime]

I have no idea why I asked that, its obviously not working as it should.

[outAtime]

For now Im just trying different combinations of [x][x]. Ill post again if anything works. Also, the engine was clearing the PV before beginning a new search, I have disabled that for now.

[Sven]

For now Im just trying different combinations of [x][x]. Ill post again if anything works. Also, the engine was clearing the PV before beginning a new search, I have disabled that for now.

If everything else regarding your implementation of pondering and pv handling were correct then using pv[0][1] as second move of the PV should do it, unless the pv[][] table has already been cleared prior to that point.

But since you report that it caused new problems I would expect that you have another bug. That is all I can say at the moment.

Sven

[outAtime]

Yes. Something else is not working properly. Ponder is calling think(), so Im guessing I need to change something there, as well as maybe a couple other places in search() where I give the engine instructions on how to handle input from the GUI.

[LiquidNitrogenOverclocker]

It sounds like this is a great time to use the printf() call.



Go through your PV and format your array (and call a conversion routine as needed) and just examine its contents.

It's much better than the aggravation of trying to remember which way you indexed your 2D array.