Andscacs - New version 0.86

[cdani]

Hi Daniel,
when I setup a position inside Shredder Chess - e.g.

4k3/3pp3/8/8/3p4/8/4N3/4KB2 b - -

As Peter told, can you try with a corrected fen like this one?

4k3/3pp3/8/8/3p4/8/4N3/4KB2 b - - 0 1

Thanks!

[cdani]

Vael Jean-Paul (jpqy) asked for a BMI2 compile. I have just done one but after 10000 games seems a little worse I will let the test go on at least to 20000, but for the moment I don't see the supposed maybe 10 elo gain.

I saw here:
https://chessprogramming.wikispaces.com ... tions-PEXT
and then the source code of Stockfish. I see that in it the gain should come from using _pext_u64 in search for magic index. Is just what I have done. Then the compile was a pgo one exactly like I do always. Someone has an idea of what is happening? Thanks!

[Werner]

Hi Daniel,
it is not my fen, it is this line:

### 4.: to Andscacs 0.86 x64 1CPU (0): position fen 4k3/3pp3/8/8/3p4/8/4N3/4KB2 b - - moves

Looks like this is sent by the GUI. Does Andscacs understand this correct?

This is the line Arena sends:
-->1:position fen 4k3/3pp3/8/8/3p4/8/4N3/4KB2 b - - 0 1

regards
Werner

PS
version 0.71 has no problems here:

Engine: Andscacs 0.71 x64 (256 MB)
von Daniel Jose
UCI debug mode is now on
### 1.: to Andscacs 0.71 x64 (0): debug on

### 2.: to Andscacs 0.71 x64 (0): ucinewgame

### 3.: to Andscacs 0.71 x64 (0): isready

### from Andscacs 0.71 x64 (0): readyok

### 4.: to Andscacs 0.71 x64 (0): position fen 4k3/3pp3/8/8/3p4/8/4N3/4KB2 w - -

### 5.: to Andscacs 0.71 x64 (0): go movestogo 40 wtime 480000 btime 480000

### from Andscacs 0.71 x64 (0): info depth 1 seldepth 4 score cp 352 nodes 26 nps 26000 time 1 pv e2d4 d7d5

1/04 0:00 +3.52 1.Sxd4 d5 (26) 26

[cdani]

Hi Daniel,
it is not my fen, it is this line:

### 4.: to Andscacs 0.86 x64 1CPU (0): position fen 4k3/3pp3/8/8/3p4/8/4N3/4KB2 b - - moves

Looks like this is sent by the GUI. Does Andscacs understand this correct?

This is the line Arena sends:
-->1:position fen 4k3/3pp3/8/8/3p4/8/4N3/4KB2 b - - 0 1

regards
Werner

I installed Schreeder GUI and I can confirm is due to lack of halfmove clock and fullmove number. I can solve this. I tell something.

[Werner]

Sorry,
it is only when it is blacks move - with white to move : no problems

[cdani]

I just updated the zip file with a version that is able to work with Schredder bug. Can you try it?

www.andscacs.com/andscacs086.zip

Thanks.

[Werner]

I just updated the zip file with a version that is able to work with Schredder bug. Can you try it?
www.andscacs.com/andscacs086.zip
Thanks.

Sorry - with black to move there is no difference:
UCI debug mode is now on
### 1.: to Andscacs 0.86b x64 1CPU (0): debug on

### 2.: to Andscacs 0.86b x64 1CPU (0): position fen 4k3/3pp3/8/8/3p4/8/4N3/4KB2 b - - moves

### 3.: to Andscacs 0.86b x64 1CPU (0): go movestogo 40 wtime 480000 btime 480000

### from Andscacs 0.86b x64 1CPU (0): info nodes 0 nps 0 time 1

### from Andscacs 0.86b x64 1CPU (0): bestmove (none)

### illegal move from engine: (none) ###
Bester Zug: O-O Zeit: 0:00.016 min

but Andscacs 0.71 works

### 11.: to Andscacs 0.71 x64 (0): ucinewgame

### 12.: to Andscacs 0.71 x64 (0): isready

### from Andscacs 0.71 x64 (0): readyok

### 13.: to Andscacs 0.71 x64 (0): position fen 4k3/3pp3/8/8/3p4/8/4N3/4KB2 b - - moves

### 14.: to Andscacs 0.71 x64 (0): go movestogo 40 wtime 480000 btime 480000

### from Andscacs 0.71 x64 (0): info depth 1 seldepth 1 score cp -168 nodes 1 nps 1000 time 1 pv d4d3

1/01 0:00 +1.68 1...d3 (1) 1
### from Andscacs 0.71 x64 (0): info depth 2 seldepth 4 score cp -217 nodes 34 nps 34000 time 1 pv d4d3 e2f4

[cdani]

Sorry - with black to move there is no difference:

I see. It depends on how do you setup the position also.
A new updated version that should work for this case:

www.andscacs.com/andscacs086.zip

[Jamal Bubker]

Thanks Daniel

[Werner]

Sorry - with black to move there is no difference:

I see. It depends on how do you setup the position also.
A new updated version that should work for this case:
www.andscacs.com/andscacs086.zip

Yes, now it works,
thanks Daniel

best wishes
Werner