Progress On Bricabrac

[Pio]

I have discovered that the "impossible to find bug" does not happen in debug mode.
It only happens in release mode.
It happens even if all optimizations are disabled.

All this time I thought I had become too stupid to logically think correctly.
But, it is not me.
It's MicroSoft!

Now I don't know what I am going to do

If anyone wants to help it happens at depth one in this position.
[d]7K/8/7p/1k5P/4P1P1/2P5/B7/8 w - - 16 64
FEN: 7K/8/7p/1k5P/4P1P1/2P5/8/1B6 w - - 0 65

Bricabrac:
1 00:00 16 16 +M0 a2b1
2 00:00 146 146 +M0 a2b1
3 00:00 1k 1k +M0 a2b1
4 00:00 11k 11k +M0 a2b1
5 00:00 36k 36k +M0 a2b1
6 00:00 37k 37k +M0 a2b1
7 00:00 38k 38k +M0 a2b1
8 00:00 44k 44k +M0 a2b1
9 00:00 58k 58k +M0 a2b1
10 00:00 97k 97k +M0 a2b1
11 00:00 187k 187k +M0 a2b1
12 00:00 435k 43,496k +M0 a2b1
13 00:00 998k 24,939k +M0 a2b1

It first shows up in this position.
[d]8/8/P2k3p/7P/4PKP1/1BP5/8/8 b - - 0 53
FEN: 8/8/P2k3p/7P/4PKP1/1BP5/8/8 b - - 0 53

Bricabrac:
1 00:00 8 8 -8.24 d6c5
2 00:00 59 59 -9.08 d6c5 a6a7
3 00:00 351 351 -9.40 d6c6 a6a7 c6b7
4 00:00 1k 1k -8.88 d6c6 f4e5 c6b6 b3c4
5 00:00 3k 3k -8.64 d6c6 f4e5 c6b6 b3c4 b6c5
6 00:00 8k 8k -9.00 d6c6 f4e5 c6c7 e5e6 c7b6 b3c4
7 00:00 14k 14k -9.00 d6c6 f4e5 c6b6 b3c4 b6a7 e5e6 a7b6
8 00:00 53k 53k -9.35 d6c6 b3c4 c6b6 f4f5 b6a7 f5g6 a7b6 g6h6
9 00:00 77k 77k -14.64 d6c6 a6a7 c6b7 e4e5 b7a7 e5e6 a7b6 e6e7 b6c5
10 00:00 711k 23,688k -M0 d6c6 a6a7 c6b7 e4e5 b7a7 e5e6 a7b6 e6e7 b6c5

Do you get any warnings when you compile the project? If so, fix the warnings. I would guess that there might be some uninitialised variables that might cause the problem.

If you just look at the position it is very seldom the white king is in the upper corner so it might have to do with that it might be the last bit or that maybe you did the same error I once made only generating moves from squares from 1 to 63 (and not from 0 to 63).

Good luck finishing off the bug!

[Mike Sherwin]

Just a stupid thought ... The debug version might differ from the release version through #ifdef or assert(...) or something like that ...

Hi Sven,

There are no #ifdef directives and no current assert(...).

[Mike Sherwin]

Begin Depth 12
12 -1512 188 6571306 d6c6 e4e5 c6b6 e5e6 b6c6 a6a7 c6b7 e6e7 b7a7 e7e8q╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠♦ a7b6 f4e5

Begin Depth 13
13 -1536 642 21469963 d6c6 e4e5 c6b6 e5e6 b6c6 f4e5 c6c7 b3d5 c7b6 e6e7 b6a6 e5e6 a6b6

Begin Depth 14
14 -1555 2610 91415295 d6c6 e4e5 c6b6 e5e6 b6c6 f4e5 c6c7 b3d5 c7b6 e6e7 b6a6 e7e8q╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠╠♦ a6b6 e8e7

Promotion related?

Probably!

[Mike Sherwin]

I have discovered that the "impossible to find bug" does not happen in debug mode.
It only happens in release mode.
It happens even if all optimizations are disabled.

All this time I thought I had become too stupid to logically think correctly.
But, it is not me.
It's MicroSoft!

Now I don't know what I am going to do

If anyone wants to help it happens at depth one in this position.
[d]7K/8/7p/1k5P/4P1P1/2P5/B7/8 w - - 16 64
FEN: 7K/8/7p/1k5P/4P1P1/2P5/8/1B6 w - - 0 65

Bricabrac:
1 00:00 16 16 +M0 a2b1
2 00:00 146 146 +M0 a2b1
3 00:00 1k 1k +M0 a2b1
4 00:00 11k 11k +M0 a2b1
5 00:00 36k 36k +M0 a2b1
6 00:00 37k 37k +M0 a2b1
7 00:00 38k 38k +M0 a2b1
8 00:00 44k 44k +M0 a2b1
9 00:00 58k 58k +M0 a2b1
10 00:00 97k 97k +M0 a2b1
11 00:00 187k 187k +M0 a2b1
12 00:00 435k 43,496k +M0 a2b1
13 00:00 998k 24,939k +M0 a2b1

It first shows up in this position.
[d]8/8/P2k3p/7P/4PKP1/1BP5/8/8 b - - 0 53
FEN: 8/8/P2k3p/7P/4PKP1/1BP5/8/8 b - - 0 53

Bricabrac:
1 00:00 8 8 -8.24 d6c5
2 00:00 59 59 -9.08 d6c5 a6a7
3 00:00 351 351 -9.40 d6c6 a6a7 c6b7
4 00:00 1k 1k -8.88 d6c6 f4e5 c6b6 b3c4
5 00:00 3k 3k -8.64 d6c6 f4e5 c6b6 b3c4 b6c5
6 00:00 8k 8k -9.00 d6c6 f4e5 c6c7 e5e6 c7b6 b3c4
7 00:00 14k 14k -9.00 d6c6 f4e5 c6b6 b3c4 b6a7 e5e6 a7b6
8 00:00 53k 53k -9.35 d6c6 b3c4 c6b6 f4f5 b6a7 f5g6 a7b6 g6h6
9 00:00 77k 77k -14.64 d6c6 a6a7 c6b7 e4e5 b7a7 e5e6 a7b6 e6e7 b6c5
10 00:00 711k 23,688k -M0 d6c6 a6a7 c6b7 e4e5 b7a7 e5e6 a7b6 e6e7 b6c5

Do you get any warnings when you compile the project? If so, fix the warnings. I would guess that there might be some uninitialised variables that might cause the problem.

If you just look at the position it is very seldom the white king is in the upper corner so it might have to do with that it might be the last bit or that maybe you did the same error I once made only generating moves from squares from 1 to 63 (and not from 0 to 63).

Good luck finishing off the bug!

Only one warning.
Severity Code Description Project File Line Suppression State
Warning C26812 The enum type '<unnamed-enum-OO>' is unscoped. Prefer 'enum class' over 'enum' (Enum.3). Bricabrac D:\source\repos\Bricabrac\Bricabrac.cpp 2282

And it does have something to do with promotion.
case 'n':
move.type = wtm ? Wn : Bn;
break;
Thanks!

[Mike Sherwin]

Replacing

enum {
OO, WP, WN, WB, WR, WRC, WQ, WK, WC, BP, BN, BB, BR, BRC, BQ, BK, BC, 16
Wd, We, Wb, Wr, Wn, Wq, Bd, Be, Bb, Br, Bn, Bq, WS, WL, BS, BL

With

constexpr auto OO = 0;
constexpr auto WP = 1;
constexpr auto WN = 2;
constexpr auto WB = 3;
constexpr auto WR = 4;
constexpr auto WRC = 5;
constexpr auto WQ = 6;
constexpr auto WK = 7;
constexpr auto WC = 8;
constexpr auto BP = 9;
constexpr auto BN = 10;
constexpr auto BB = 11;
constexpr auto BR = 12;
constexpr auto BRC = 13;
constexpr auto BQ = 14;
constexpr auto BK = 15;
constexpr auto BC = 16;
constexpr auto Wd = 17;
constexpr auto We = 18;
constexpr auto Wb = 19;
constexpr auto Wr = 20;
constexpr auto Wn = 21;
constexpr auto Wq = 22;
constexpr auto Bd = 23;
constexpr auto Be = 24;
constexpr auto Bb = 25;
constexpr auto Br = 26;
constexpr auto Bn = 27;
constexpr auto Bq = 28;
constexpr auto WS = 29;
constexpr auto WL = 30;
constexpr auto BS = 31;
constexpr auto BL = 32;

Got rid of the warning but did not change the behavior. So the bug must be in the promotion code.

[Mike Sherwin]

Despite the only warning having something to do with promotion and the spurious characters in the pv it was a "red herring". The spurious characters were caused by me not terminating the move string after appending the promotion type.

Code: Select all

      xbmove[0] = 'a' + (pvs[0][i].fs & 7);
      xbmove[1] = '1' + (pvs[0][i].fs >> 3);
      xbmove[2] = 'a' + (pvs[0][i].ts & 7);
      xbmove[3] = '1' + (pvs[0][i].ts >> 3);
      xbmove[4] = '\0';
//      if (pvs[0][i].type == Wq || pvs[0][i].type == Bq) { xbmove[4] = 'q'; xbmove[5] = '\0'; }
//      vs.
//      if (pvs[0][i].type == Wq || pvs[0][i].type == Bq) xbmove[4] = 'q'; 
      

So now it is back to Bricabrac running fine in debug mode and flaking out in release mode.

Arena running the debug build.
FEN: 8/P7/3k3p/7P/4PpP1/1BP2K2/8/8 b - - 0 53

Bricabrac:
1 00:00 19 19 +8.24 f3f4
2 00:00 98 98 +14.09 a6a7 d6e5
3 00:00 467 467 +14.09 a6a7 d6e5 a7a8q
4 00:00 2k 2k +15.21 a6a7 d6e5 a7a8q e5d6
5 00:00 8k 8k +15.21 a6a7 d6e5 a7a8q e5d6 f3f4
6 00:00 14k 14k +15.80 a6a7 d6d7 a7a8q d7e7 a8d5 e7f6
7 00:00 29k 29k +15.75 a6a7 d6e5 a7a8q e5d6 f3f4 d6c5 a8d5
8 00:00 62k 6,205k +15.95 a6a7 d6e5 a7a8q e5f6 f3f4 f6g7 a8d8 g7h7
9 00:00 191k 3,816k +15.91 a6a7 d6e7 f3f4 e7f6 b3c4 f6e7 a7a8q e7d6 a8d5
10 00:00 619k 3,643k +16.56 a6a7 d6e5 a7a8q e5d6 a8d5 d6c7 f3f4 c7b6 d5d6 b6b5

Arena running the release build.
FEN: 8/8/P2k3p/7P/4PKP1/1BP5/8/8 b - - 0 53

Bricabrac:
1 00:00 8 8 -8.24 d6c5
2 00:00 59 59 -9.08 d6c5 a6a7
3 00:00 352 352 -9.40 d6c6 a6a7 c6b7
4 00:00 1k 1k -8.88 d6c6 f4e5 c6b6 b3c4
5 00:00 3k 3k -8.64 d6c6 f4e5 c6b6 b3c4 b6c5
6 00:00 8k 8k -9.00 d6c6 f4e5 c6c7 e5e6 c7b6 b3c4
7 00:00 14k 14k -9.00 d6c6 f4e5 c6b6 b3c4 b6a7 e5e6 a7b6
8 00:00 53k 53k -9.35 d6c6 b3c4 c6b6 f4f5 b6a7 f5g6 a7b6 g6h6
9 00:00 83k 83k -14.64 d6c6 a6a7 c6b7 e4e5 b7a7 e5e6 a7b6 e6e7 b6c5
10 00:00 718k 23,931k -M0 d6c6 a6a7 c6b7 e4e5 b7a7 e5e6 a7b6 e6e7 b6c5

So now what do I do? I can't afford learning a new compiler because of my learning disability and the time that would take me. Please, can someone verify the compiler (MSVS 2019 Community) issue and find a solution?
https://www.mediafire.com/file/zm7tbv8o ... c.zip/file

[Mike Sherwin]

I goofed that up. Let me try again.

Arena running the debug build.
FEN: 8/8/P1k4p/7P/4PKP1/1BP5/8/8 w - - 1 54

Bricabrac:
1 00:00 8 8 -8.24 d6c5
2 00:00 59 59 -9.08 d6c5 a6a7
3 00:00 352 352 -9.40 d6c6 a6a7 c6b7
4 00:00 1k 1k -8.88 d6c6 f4e5 c6b6 b3c4
5 00:00 3k 3k -8.64 d6c6 f4e5 c6b6 b3c4 b6c5
6 00:00 8k 8k -9.00 d6c6 f4e5 c6c7 e5e6 c7b6 b3c4
7 00:00 14k 14k -9.00 d6c6 f4e5 c6b6 b3c4 b6a7 e5e6 a7b6
8 00:00 55k 5,510k -9.35 d6c6 b3c4 c6b6 f4f5 b6a7 f5g6 a7b6 g6h6
9 00:00 85k 4,255k -14.64 d6c6 a6a7 c6b7 e4e5 b7a7 e5e6 a7b6 e6e7 b6c5
10 00:00 373k 3,728k -11.10 d6c6 b3c4 c6b6 e4e5 b6c6 e5e6 c6c7 a6a7 c7b7 e6e7

Arena running the release build.
FEN: 8/8/P2k3p/7P/4PKP1/1BP5/8/8 b - - 0 53

Bricabrac:
1 00:00 8 8 -8.24 d6c5
2 00:00 59 59 -9.08 d6c5 a6a7
3 00:00 352 352 -9.40 d6c6 a6a7 c6b7
4 00:00 1k 1k -8.88 d6c6 f4e5 c6b6 b3c4
5 00:00 3k 3k -8.64 d6c6 f4e5 c6b6 b3c4 b6c5
6 00:00 8k 8k -9.00 d6c6 f4e5 c6c7 e5e6 c7b6 b3c4
7 00:00 14k 14k -9.00 d6c6 f4e5 c6b6 b3c4 b6a7 e5e6 a7b6
8 00:00 53k 53k -9.35 d6c6 b3c4 c6b6 f4f5 b6a7 f5g6 a7b6 g6h6
9 00:00 83k 83k -14.64 d6c6 a6a7 c6b7 e4e5 b7a7 e5e6 a7b6 e6e7 b6c5
10 00:00 718k 23,931k -M0 d6c6 a6a7 c6b7 e4e5 b7a7 e5e6 a7b6 e6e7 b6c5

[Mike Sherwin]

Bricabrac debug build vs TSCP181
[pgn][Event "TSCP.at"]
[Site "DESKTOP-HFVHK2B"]
[Date "2021.04.28"]
[Round "3"]
[White "Bricabrac"]
[Black "Tscp181"]
[Result "1-0"]
[BlackElo "2200"]
[ECO "A90"]
[Opening "Dutch"]
[Time "19:10:33"]
[Variation "2.c4 Nf6 3.g3 e6 4.Bg2 d5"]
[WhiteElo "2000"]
[TimeControl "5+0"]
[Termination "normal"]
[PlyCount "115"]
[WhiteType "human"]
[BlackType "human"]

1. d4 f5 2. c4 Nf6 3. g3 e6 4. Bg2 d5 {(d7d5 g1f3 d5c4 d1a4 b8c6 a4c4)
+0.41/5} 5. cxd5 {(c4d5 f8b4 c1d2 b4d2 b1d2 e6d5 g1f3 b8c6) +0.50/8} exd5
{(e6d5 b1c3 b8c6 g1f3 f8d6) +0.57/5} 6. Nf3 {(g1f3 c7c5 b1c3 b8c6 c1e3 c5d4
f3d4 h7h6) +0.47/8} Nc6 {(b8c6 b1c3 f8b4 e1g1 b4c3 b2c3) +0.55/5} 7. O-O
{(e1g1 c8e6 b1c3 f8e7 c1f4 h7h6 d1b3 b7b6) +0.76/8} Bd6 {(f8d6 b1c3 e8g8
f3e5 c6e5 d4e5 d6e5 g2d5) +0.52/5} 8. Nc3 {(b1c3 c8d7 c1e3 h8f8 a2a3 h7h6
d1b3 c6a5) +0.82/8} O-O {(e8g8 c1g5 c8e6 e2e3 f8e8 g5f6 d8f6) +0.45/5} 9.
Nb5 {(c3b5 d6b4 c1f4 b4a5 e2e3 c8e6 h2h3) +0.86/7} Be6 {(c8e6 b5d6 d8d6
c1f4 d6e7) +0.57/5} 10. Ng5 {(f3g5 e6f7 b5d6 d8d6 g5f7 f8f7 e2e3) +1.12/7}
Re8 {(f8e8 c1e3 a7a5 b5d6 d8d6 g5e6 e8e6) +0.69/4} 11. Nxd6 {(b5d6 d8d6
c1f4 d6d8 g5e6 e8e6 e2e3) +1.34/7} Qxd6 {(d8d6 c1f4 d6b4 e2e3 b4b2 g5e6
e8e6 f4c7) +0.50/5} 12. Bf4 {(c1f4 d6d7 g5e6 e8e6 a1c1 a8d8 h2h3 h7h6)
+1.08/8} Qb4 {(d6b4 e2e3 b4b2 g5e6 e8e6 f4c7) +0.50/5} 13. Nxe6 {(g5e6 e8e6
e2e3 b4e7 d1c2 f6e4 a1d1) +1.53/7} Rxe6 {(e8e6 e2e3 b4b2 d1d3 e6e7 d3f5)
+0.63/5} 14. e3 {(e2e3 b4b6 d1c2 f6e4 f2f3 c6b4 c2c7) +1.38/7} Qxb2 {(b4b2
f4c7 a8c8 a1b1) +0.55/4} 15. Rb1 {(a1b1 b2a2 b1b7 a8c8 g2h3 g7g6 f4c7 e6e4)
+1.45/8} Qxa2 {(b2a2 b1b7 a7a5 g2f3 a8a7 b7a7 c6a7 f4c7) +0.52/5} 16. Rxb7
{(b1b7 a8c8 b7c7 c8c7 f4c7 f6e4 c7f4) +1.63/7} a5 {(a7a5 b7c7 a5a4 d1b1
a2b1 f1b1) +0.38/5} 17. Rxc7 {(b7c7 a8d8 h2h3 h7h6 g3g4 f6e8 c7b7) +2.01/7}
a4 {(a5a4 d1d3 c6e7 f4e5 f6e4) +0.48/5} 18. Qc1 {(d1c1 c6b4 c1c3 a2b3 c3b3
a4b3 f4e5) +2.04/7} Ne7 {(c6e7 f4g5 e7g6 g5f6 e6f6) +0.38/5} 19. Bg5 {(f1d1
e7g6 c7c8 e6e8 c8a8 e8a8 f4d6) +1.90/7} Kf7 {(g8f7 g5f6 f7f6 c1b1 a2b1
f1b1) +0.01/5} 20. Bxf6 {(g5f6 f7f6 c1d1 a8d8 h2h3 e7g6 f1e1) +2.30/7} Kxf6
{(f7f6 c1c5 a8d8 f1c1 a4a3) -0.02/5} 21. Rd7 {(c7d7 a8e8 c1d1 f6g6 h2h3
a2b3 d1d2) +2.22/7} Ra5 {(a8a5 c1c3 a5b5 f1a1 a2b3) +0.12/5} 22. Qc7 {(c1d1
f6g6 h2h4 h7h6 h4h5 g6h7 f1e1) +2.54/7} Qd2 {(a2d2 c7d8 f6f7 g2d5 a5d5 d7d5
e7d5 d8d5) -0.91/5} 23. Qd8 {(c7d8 f6f7 d7c7 f5f4 g2d5 a5d5 d8d5 f4e3)
+4.20/8} Qb4 {(d2b4 g2d5 e6b6 d8f8 f6g6 d7e7 a5d5 e7g7) -2.64/5} 24. Bxd5
{(g2d5 a5d5 d7d5 b4c3 d5d7 f6f7 d8h8 f7g6) +4.96/8} Rxd5 {(a5d5 d7d5 f6f7
d8d7 e6f6 f1c1) -3.24/6} 25. Rxd5 {(d7d5 b4c3 d5d7 c3c8 d8c8 e7c8 d4d5
e6e4) +5.18/8} Kf7 {(f6f7 d8d7 e6f6 d5b5 b4d6) -3.24/5} 26. Rd7 {(d5d7 e6e4
d4d5 f7g6 d8e8 g6h6 e8f7 a4a3) +5.34/8} Kg6 {(f7g6 d8a8 b4b5 d7a7 b5d5 a7a4
d5a8 a4a8) -4.23/5} 27. Qe8+ {(d8e8 g6h6 d4d5 e6e4 d5d6 e7g6 e8f7 b4c3)
+5.58/8} Kh6 {(g6h6 d7a7 e6e4 e8a8 a4a3 a7a3) -4.89/5} 28. Qf7 {(e8f7 e6e4
d4d5 b4c4 d7e7 e4e7 f7e7 c4d5) +5.82/8} Re4 {(e6e4 f1a1 f5f4 g3f4 a4a3
d4d5) -4.71/6} 29. d5 {(d4d5 g7g6 d5d6 e7c6 f7h7 h6g5 h2h4 g5f6) +6.19/8}
Qb5 {(b4b5 d7e7 e4e7 f7e7 b5d5 f1b1) -4.54/6} 30. Rxe7 {(d7e7 e4e7 f7e7
b5d5 e7h4 h6g6 h4a4 d5f7 e3e4) +6.48/9} Rxe7 {(e4e7 f7e7 b5d5 e7h4 h6g6
h4a4 g6f7) -5.93/6} 31. Qxe7 {(f7e7 b5d5 e7h4 h6g6 h4a4 h7h5 f1d1 d5e4
a4e4) +5.61/9} Qxd5 {(b5d5 f1a1 d5e4 e7d7 f5f4 a1a4) -6.14/6} 32. Ra1
{(f1a1 h6g6 e7e8 g6f6 h2h4 h7h5 a1a4 g7g6 g1f1) +5.61/9} Kg6 {(h6g6 e7a7
d5f3 a1a4 f3d1) -6.17/5} 33. Qb4 {(e7b4 g6f6 a1a4 f6e6 b4d4 d5d4 a4d4 e6e5)
+5.58/8} Qe4 {(d5e4 b4b5 a4a3 a1a3 e4e7 a3a2) -6.17/6} 34. Rxa4 {(a1a4 e4d3
b4d4 d3d4 a4d4 g6f6 g1g2 f6e5 g2f3) +5.89/9} Qxb4 {(e4b4 a4b4 g6f7 b4b7
f7g6 f2f3 g6f6) -5.86/6} 35. Rxb4 {(a4b4 g6f6 b4b7 h7h6 g1g2 h6h5 h2h4 f6g6
g2f3 g6f6 f3f4) +6.11/11} Kf7 {(g6f7 b4b7 f7f6 f2f3 h7h6 e3e4 g7g6 g1f2)
-6.21/7} 36. Rb6 {(g1f1 f7f6 f1e2 h7h5 b4b7 g7g5 e2d3 f6e5 d3c4) +6.08/9}
Ke7 {(f7e7 f2f3 h7h5 e3e4 f5e4 b6b7 e7e6 f3e4) -6.43/7} 37. Kg2 {(g1g2 h7h6
g2f3 e7f7 f3f4 g7g6 f4e5 h6h5 b6b7 f7f8) +6.62/10} Kf7 {(e7f7 g2f3 h7h6
g3g4 f5g4 f3g4 g7g6 e3e4) -6.62/7} 38. Kf3 {(g2f3 g7g6 b6b7 f7e6 b7h7 g6g5
h7c7 e6e5 c7c5 e5e6) +6.94/10} Ke7 {(f7e7 g3g4 f5g4 f3g4 e7f7 e3e4 f7e7
e4e5) -6.79/7} 39. Kf4 {(f3f4 g7g6 f4e5 e7e8 b6b7 h7h5 e5f6 e8d8 f6g6 d8c8)
+7.48/10} g6 {(g7g6 g3g4 f5g4 b6b7 e7f8 b7h7 f8g8 h7d7) -7.19/7} 40. Ke5
{(f4e5 e7d7 b6b7 d7c6 b7g7 c6c5 g7h7 c5c4 h7g7 f5f4) +8.03/10} Kd7 {(e7d7
b6b7 d7c6 b7h7 c6b6 h7h6 b6c5 h6g6) -8.20/7} 41. Rb7+ {(b6b7 d7c6 b7h7 c6b5
h7h6 g6g5 e5f5 b5c4 h6g6 g5g4) +8.86/10} Kc6 {(d7c6 b7h7 c6c5 h7g7 c5b5
g7g6 b5c5 e5f5) -9.96/7} 42. Rxh7 {(b7h7 c6c5 h7g7 f5f4 g3f4 c5c4 g7g6 c4c3
g6c6 c3b4) +9.03/10} Kc5 {(c6c5 h7g7 c5c4 g7g6 c4d3 g6g5 d3c4 g5f5)
-10.06/7} 43. Rg7 {(h7g7 f5f4 e5f4 c5d6 f4e4 d6c5 g7g6 c5b5 g6g7 b5c4)
+8.99/10} Kc4 {(c5c4 g7g6 c4d3 g6f6 d3c4 f6f5 c4c5 h2h4) -10.56/8} 44. Rxg6
{(g7g6 f5f4 g6g4 c4c3 g4f4 c3d2 f4f8 d2c3 e5e6 c3c4) +9.25/10} Kd3 {(c4d3
g6f6 f5f4 f6f4 d3c3 h2h4 c3d3 e3e4) -10.94/8} 45. Rf6 {(g6f6 f5f4 f6f4 d3c3
e5e6 c3d2 f4f8 d2c3 f8c8 c3b4) +9.33/10} f4 {(f5f4 f6f4 d3e2 e3e4 e2d3 h2h4
d3c4 f2f3) -11.25/8} 46. Rxf4 {(f6f4 d3c3 h2h4 c3d2 h4h5 d2c3 h5h6 c3d2
h6h7 d2c3) +17.23/10} Ke2 {(d3e2 e3e4 e2d3 f4f7 d3c3 h2h4 c3d3 f2f4)
-11.60/8} 47. h4 {(h2h4 e2d3 h4h5 d3c3 h5h6 c3d2 h6h7 d2d3 e5e6 d3c3)
+17.35/10} Kd2 {(e2d2 h4h5 d2d3 h5h6 d3e2 h6h7 e2e1 h7h8q) -18.66/8} 48. h5
{(h4h5 d2c3 h5h6 c3d2 h6h7 d2c3 h7h8q c3d2 f4c4 d2e2) +17.43/10} Kd3 {(d2d3
h5h6 d3e2 h6h7 e2d3 e3e4 d3e2 h7h8q) -19.04/8} 49. h6 {(h5h6 d3c3 h6h7 c3d2
f4f8 d2c3 h7h8q c3c4 h8h5 c4c3) +17.57/10} Kd2 {(d3d2 h6h7 d2e2 e3e4 e2e1
h7h8q e1e2 f2f3) -19.35/8} 50. h7 {(h6h7 d2c3 e5e6 c3d2 f4f8 d2d3 h7h8q
d3e4 h8e5) +17.87/9} Ke2 {(d2e2 e3e4 e2e1 h7h8q e1e2 f4f7 e2e1 f2f4)
-19.70/8} 51. g4 {(g3g4 e2f1 f4f8 f1g2 g4g5 g2h3 h7h8q h3g4 g5g6) +18.21/9}
Ke1 {(e2e1 e3e4 e1d2 h7h8q d2d3 f4f7 d3e2 f2f4) -20.00/8} 52. g5 {(g4g5
e1d2 g5g6 d2c3 g6g7 c3d2 g7g8q d2c3 h7h8q) +25.34/9} Ke2 {(e1e2 h7h8q e2d2
g5g6 d2d3 g6g7 d3e2 g7g8q) -26.39/8} 53. g6 {(g5g6 e2d2 g6g7 d2c3 g7g8q
c3d2 h7h8q d2c3 g8g5) +25.81/9} Ke1 {(e2e1 h7h8q e1e2 g6g7 e2d2 g7g8q d2e2
e3e4) -26.77/8} 54. g7 {(g6g7 e1d2 g7g8q d2c3 g8g5 c3d2 h7h8q d2c3 e5e6)
+25.93/9} Kd2 {(e1d2 g7g8q d2e2 g8g1 e2d2 g1f1 d2c3 h7h8q) -27.05/7} 55.
g8=Q {(g7g8q d2c3 g8g5 c3d2 f4f8 d2c3 h7h8q c3c4 e5e6) +26.10/9} Ke1 {(d2e1
g8g1 e1e2 e3e4 e2d2 g1f1 d2c3 h7h8q) -27.43/6} 56. Qa2 {(g8a2 e1f1 f4g4
f1e1 g4g1 c3c4 h8h5 c4c3) +M3/8} Kf1 {(e1f1 f4g4 f1e1 g4g1) -M2/5} 57. Rg4
{(f4g4 f1e1 g4g1 e1e2) +M2/8} Ke1 {(f1e1 g4g1) -M1/3} 58. Rg1# {(g4g1 e1d2)
+M1/8} 1-0
[/pgn]

[lithander]

One thing you could try is that you get rid of all the uninitialized variables by initializing them or make extra sure that they really really get initialized before use. That's one of the differences between release and debug-mode: Afaik in debug mode the memory get's initialized with magic debug values while in release modes all bets are off. The code ceases to run deterministically because each time you start your uninitialized variables can have a different value. A debuggin nightmare.

[Mike Sherwin]

One thing you could try is that you get rid of all the uninitialized variables by initializing them or make extra sure that they really really get initialized before use. That's one of the differences between release and debug-mode: Afaik in debug mode the memory get's initialized with magic debug values while in release modes all bets are off. The code ceases to run deterministically because each time you start your uninitialized variables can have a different value. A debuggin nightmare.

Thank you Thomas! However, there are no uninitialized variables that are not first assigned to before use. If there were the compiler would flag it with a warning. However, I will initialize every variable just to double check.