TalkChess.com

Computer Chess Club
https://talkchess.com/

Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems

https://talkchess.com/viewtopic.php?t=77438

Page 27 of 29

Re: Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems

Posted: Thu Jul 29, 2021 2:53 am
by AndrewGrant
Graham Banks wrote: ↑Thu Jul 29, 2021 2:37 am
supersharp77 wrote: ↑Thu Jul 29, 2021 1:13 am
Graham Banks wrote: ↑Thu Jul 29, 2021 12:15 am Has anybody had success using Ethereal 13.00 in the Shredder GUI?
Time issues In Fritz Gui (Like almost all the other Versions) What problems is it having in the Shredder GUI? I don't recall time losses in Banksia or Chess OK.....Thx AR :) :wink:
No problems in ChessGUI, but is losing all of its games under Shredder.
Examples below. It's lost 10 games straight.

[pgn][Event "40 Moves in 16 min"]
[Site "Revenge Gauntlet, GRAHAM-Z87"]
[Date "2021.07.25"]
[Round "1.3"]
[White "Revenge 1.0 64-bit"]
[Black "Ethereal 13.00 64-bit"]
[ECO "E04"]
[Result "1-0"]

1. d4 {book 0s} d5 {book 0s} 2. Nf3 {book 0s} e6 {book 0s}
3. g3 {book 0s} Nf6 {book 0s} 4. c4 {book 0s} dxc4 {book
0s} 5. Bg2 {book 0s} Bb4+ {book 0s} 6. Bd2 {book 0s} a5
{book 0s} 7. Qc2 {book 0s} Bxd2+ {book 0s} 8. Qxd2 {book
0s} c6 {book 0s} 9. a4 {book 0s} Ne4 {-0.20/23 55s} 10. Qf4
{+0.57/31 34s} Qb6 {-0.20/20 17s (Nd6)} 11. Ne5 {+0.29/27
26s (Qc1)} Nd6 {-0.20/21 23s (Qb4+)} 12. O-O {+0.89/32 29s
(Qc1)} Qb4 {-0.20/23 41s (f6)} 13. Na3 {+1.08/32 40s (Qc1)}
O-O {-0.20/23 26s (f6)} 14. Rfc1 {+1.49/29 32s (g4)} f6
{+0.20/22 1:01m} 15. Nexc4 {+1.22/29 23s} Nxc4 {+0.20/24
17s (Rd8)} 16. Nxc4 {+1.90/30 27s (Rxc4)} e5 {+0.20/28 46s}
17. dxe5 {+1.95/35 1:12m} Be6 {+0.20/28 44s} 18. Nd6
{+1.95/36 28s} Qxf4 {+0.20/31 18s (Qb6)} 19. gxf4 {+3.26/32
24s} fxe5 {+0.20/33 20s} 20. fxe5 {+3.31/32 48s} g6
{+0.20/32 18s (Ra7)} 21. Nxb7 {+5.09/28 26s} Ra7 {+0.20/33
19s} 22. Nc5 {+5.11/30 47s} Bf5 {+0.20/32 18s (Bg4)}
23. Ra3 {+5.47/27 24s (Nb3)} Nd7 {+0.48/28 3:09m} 24. Re3
{+5.52/27 21s (Bxc6)} Nxc5 {+0.40/27 23s (Rb8)} 25. Rxc5
{+5.69/30 30s} Rc8 {+0.20/30 1:11m (Kg7)} 26. b3 {+6.19/28
25s (Bxc6)} Kg7 {+0.20/31 9s (Be6)} 27. Bh3 {+6.42/31 1:27m
(e6)} Bxh3 {+0.20/34 11s} 28. Rxh3 {+6.27/28 18s} Ra6
{+0.20/34 25s (Rf8)} 29. Kg2 {+6.47/29 19s (e6)} Rca8
{+0.20/35 8s (Rb8)} 30. Rd3 {+6.99/27 21s (Rhc3)} R8a7
{+0.20/32 11s (Rb8)} 31. f4 {+7.54/31 58s (e6)} Rf7
{+0.20/32 9s (Rb7)} 32. e3 {+7.68/31 53s (Kh1)} Re7
{+0.20/29 10s (Rb7)} 33. Kf3 {+7.82/31 44s (f5)} Rb7
{+0.20/33 53s (Kf7)} 34. h4 {+7.89/29 14s (Rdc3)} Kf8
{+0.20/32 7s (Kf7)} 35. Rdc3 {+8.14/27 13s (e6)} Rbb6
{+0.20/36 9s} 36. Kg4 {+8.28/27 9s (Rd3)} Kf7 {+0.20/31 7s
(Rb7)} 37. Kg5 {+8.57/25 7s (f5)} Kg7 {+3.48/27 19s (Ra7)}
38. h5 {+8.87/28 19s (Rd3)} h6+ {+5.81/27 23s (Kf7)}
39. Kg4 {+9.40/25 10s} Rb4 {+6.31/28 25s (gxh5+)} 40. Rxc6
{+11.02/28 11s (hxg6)} 1-0[/pgn]

[pgn][Event "40 Moves in 16 min"]
[Site "Revenge Gauntlet, GRAHAM-Z87"]
[Date "2021.07.25"]
[Round "2.3"]
[White "Ethereal 13.00 64-bit"]
[Black "Revenge 1.0 64-bit"]
[ECO "B15"]
[Result "0-1"]

1. e4 {book 0s} c6 {book 0s} 2. Nc3 {book 0s} d5 {book 0s}
3. d4 {book 0s} dxe4 {book 0s} 4. Nxe4 {book 0s} Nf6 {book
0s} 5. Nxf6+ {book 0s} exf6 {book 0s} 6. c3 {book 0s} Bd6
{book 0s} 7. Bd3 {book 0s} O-O {book 0s} 8. Qc2 {book 0s}
h6 {book 0s} 9. Kf1 {-0.20/19 55s} c5 {-0.89/29 36s (Re8)}
10. dxc5 {+0.20/30 24s} Bxc5 {-0.88/30 52s (Bc7)} 11. Bf4
{+0.20/25 1:40m} Qb6 {-0.74/28 1:42m (Kh8)} 12. Rd1
{+0.20/28 18s (Nf3)} a5 {-1.08/27 23s (f5)} 13. b4
{+0.10/23 3:20m (Nf3)} axb4 {-4.00/28 47s (Be7)} 14. cxb4
{-0.20/25 14s} Qxb4 {-4.15/28 59s (Bxb4)} 15. Ne2 {-0.20/25
18s (Bxb8)} Nc6 {-4.45/29 30s (Bg4)} 16. Rb1 {-0.20/26 14s
(Bc4)} Qa3 {-4.30/26 20s} 17. Rb3 {-0.20/29 12s (Kg1)} Qxa2
{-6.04/27 28s} 18. Rb2 {-2.64/28 1:54m (Qxa2)} Qa1+
{-6.59/27 1:11m (Qa7)} 19. Rb1 {-1.76/30 2:00m} Qa7
{-6.31/30 17s (Qa3)} 20. Be4 {-3.76/26 41s (h3)} Re8
{-6.90/27 25s (Bxf2)} 21. Bf3 {-4.95/26 33s (Ng3)} Bxf2
{-7.08/27 36s} 22. h3 {-5.43/27 22s (h4)} Be3 {-7.20/27
17s} 23. Rd1 {-6.04/29 20s} Be6 {-7.61/27 16s (Bxf4)} 0-1[/pgn]

[pgn][Event "40 Moves in 16 min"]
[Site "Revenge Gauntlet, GRAHAM-Z87"]
[Date "2021.07.26"]
[Round "3.3"]
[White "Revenge 1.0 64-bit"]
[Black "Ethereal 13.00 64-bit"]
[ECO "A33"]
[Result "1-0"]

1. c4 {book 0s} c5 {book 0s} 2. Nf3 {book 0s} Nc6 {book 0s}
3. d4 {book 0s} cxd4 {book 0s} 4. Nxd4 {book 0s} Nf6 {book
0s} 5. Nc3 {book 0s} e6 {book 0s} 6. a3 {book 0s} Bc5 {book
0s} 7. Nb3 {book 0s} Bb6 {book 0s} 8. e4 {book 0s} O-O
{book 0s} 9. Be2 {+0.06/26 36s} a5 {+0.20/22 26s (d6)}
10. c5 {+0.43/28 44s} a4 {+0.20/26 38s (Bc7)} 11. cxb6
{+0.60/29 23s} axb3 {+0.20/28 22s} 12. Bf4 {+0.60/31 27s
(Qxb3)} e5 {+0.20/28 22s (Qxb6)} 13. Be3 {+1.14/29 50s} Ne8
{+0.20/29 34s (d6)} 14. Qxb3 {+2.94/29 22s (Bb5)} Nd4
{+0.20/32 28s (d6)} 15. Bxd4 {+3.48/34 28s} exd4 {+0.20/34
1:00m} 16. Nd5 {+3.56/33 1:24m} Nf6 {+0.20/33 17s} 17. O-O
{+3.56/34 25s (Nc7)} Nxd5 {+0.20/30 25s} 18. exd5 {+3.56/33
33s} Qf6 {+0.20/32 28s (Re8)} 19. Rfc1 {+4.29/28 29s (Bd3)}
Qd6 {+0.20/31 25s (Re8)} 20. Rc7 {+5.39/30 55s (Bd3)} g6
{+0.20/32 26s (Qe5)} 21. Rd1 {+6.36/29 1:04m (Bd3)} Kg7
{+0.20/28 1:28m (Qe5)} 22. Rxd4 {+6.73/29 23s (g3)} Qf6
{+3.56/29 1:41m (Re8)} 23. Qb4 {+6.96/26 21s (Rdc4)} d6
{+3.70/31 1:07m} 24. h3 {+7.12/28 36s (Rf4)} h5 {+3.62/27
1:01m} 25. a4 {+7.45/27 25s (Rf4)} Qe5 {+3.29/27 1:15m
(Rb8)} 26. Re4 {+7.71/24 18s} Qxd5 {+4.44/27 21s (Qg5)}
27. Bf3 {+7.81/27 24s} Be6 {+4.94/27 23s (Qg5)} 28. Re3
{+8.22/28 16s} Qa2 {+5.24/29 37s} 29. b3 {+8.09/28 17s
(Ra3)} Qa1+ {+5.83/25 31s (Bxb3)} 30. Re1 {+9.87/27 28s
(Kh2)} 1-0[/pgn]
+20 or -20 eval on every turn. It means the NNUE file is not loaded. Something I fixed and aborted on in 13.25, which is why Ray had issues initially since Cutechess was sending <empty> unexpectedly.

I believe that would imply that the Shredder GUI perhaps requires absolute file paths for the weight files instead of relative ones? Each GUI seems to be different. The advice I've given all those who have E13 is to get the absolute file path from the file explorer. Something like "C:/Users/Andrew/Desktop/MyEngines/..".

Re: Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems

Posted: Thu Jul 29, 2021 3:02 am
by Graham Banks
AndrewGrant wrote: ↑Thu Jul 29, 2021 2:53 am
Graham Banks wrote: ↑Thu Jul 29, 2021 2:37 am
supersharp77 wrote: ↑Thu Jul 29, 2021 1:13 am
Graham Banks wrote: ↑Thu Jul 29, 2021 12:15 am Has anybody had success using Ethereal 13.00 in the Shredder GUI?
Time issues In Fritz Gui (Like almost all the other Versions) What problems is it having in the Shredder GUI? I don't recall time losses in Banksia or Chess OK.....Thx AR :) :wink:
No problems in ChessGUI, but is losing all of its games under Shredder.
Examples below. It's lost 10 games straight.

[pgn][Event "40 Moves in 16 min"]
[Site "Revenge Gauntlet, GRAHAM-Z87"]
[Date "2021.07.25"]
[Round "1.3"]
[White "Revenge 1.0 64-bit"]
[Black "Ethereal 13.00 64-bit"]
[ECO "E04"]
[Result "1-0"]

1. d4 {book 0s} d5 {book 0s} 2. Nf3 {book 0s} e6 {book 0s}
3. g3 {book 0s} Nf6 {book 0s} 4. c4 {book 0s} dxc4 {book
0s} 5. Bg2 {book 0s} Bb4+ {book 0s} 6. Bd2 {book 0s} a5
{book 0s} 7. Qc2 {book 0s} Bxd2+ {book 0s} 8. Qxd2 {book
0s} c6 {book 0s} 9. a4 {book 0s} Ne4 {-0.20/23 55s} 10. Qf4
{+0.57/31 34s} Qb6 {-0.20/20 17s (Nd6)} 11. Ne5 {+0.29/27
26s (Qc1)} Nd6 {-0.20/21 23s (Qb4+)} 12. O-O {+0.89/32 29s
(Qc1)} Qb4 {-0.20/23 41s (f6)} 13. Na3 {+1.08/32 40s (Qc1)}
O-O {-0.20/23 26s (f6)} 14. Rfc1 {+1.49/29 32s (g4)} f6
{+0.20/22 1:01m} 15. Nexc4 {+1.22/29 23s} Nxc4 {+0.20/24
17s (Rd8)} 16. Nxc4 {+1.90/30 27s (Rxc4)} e5 {+0.20/28 46s}
17. dxe5 {+1.95/35 1:12m} Be6 {+0.20/28 44s} 18. Nd6
{+1.95/36 28s} Qxf4 {+0.20/31 18s (Qb6)} 19. gxf4 {+3.26/32
24s} fxe5 {+0.20/33 20s} 20. fxe5 {+3.31/32 48s} g6
{+0.20/32 18s (Ra7)} 21. Nxb7 {+5.09/28 26s} Ra7 {+0.20/33
19s} 22. Nc5 {+5.11/30 47s} Bf5 {+0.20/32 18s (Bg4)}
23. Ra3 {+5.47/27 24s (Nb3)} Nd7 {+0.48/28 3:09m} 24. Re3
{+5.52/27 21s (Bxc6)} Nxc5 {+0.40/27 23s (Rb8)} 25. Rxc5
{+5.69/30 30s} Rc8 {+0.20/30 1:11m (Kg7)} 26. b3 {+6.19/28
25s (Bxc6)} Kg7 {+0.20/31 9s (Be6)} 27. Bh3 {+6.42/31 1:27m
(e6)} Bxh3 {+0.20/34 11s} 28. Rxh3 {+6.27/28 18s} Ra6
{+0.20/34 25s (Rf8)} 29. Kg2 {+6.47/29 19s (e6)} Rca8
{+0.20/35 8s (Rb8)} 30. Rd3 {+6.99/27 21s (Rhc3)} R8a7
{+0.20/32 11s (Rb8)} 31. f4 {+7.54/31 58s (e6)} Rf7
{+0.20/32 9s (Rb7)} 32. e3 {+7.68/31 53s (Kh1)} Re7
{+0.20/29 10s (Rb7)} 33. Kf3 {+7.82/31 44s (f5)} Rb7
{+0.20/33 53s (Kf7)} 34. h4 {+7.89/29 14s (Rdc3)} Kf8
{+0.20/32 7s (Kf7)} 35. Rdc3 {+8.14/27 13s (e6)} Rbb6
{+0.20/36 9s} 36. Kg4 {+8.28/27 9s (Rd3)} Kf7 {+0.20/31 7s
(Rb7)} 37. Kg5 {+8.57/25 7s (f5)} Kg7 {+3.48/27 19s (Ra7)}
38. h5 {+8.87/28 19s (Rd3)} h6+ {+5.81/27 23s (Kf7)}
39. Kg4 {+9.40/25 10s} Rb4 {+6.31/28 25s (gxh5+)} 40. Rxc6
{+11.02/28 11s (hxg6)} 1-0[/pgn]

[pgn][Event "40 Moves in 16 min"]
[Site "Revenge Gauntlet, GRAHAM-Z87"]
[Date "2021.07.25"]
[Round "2.3"]
[White "Ethereal 13.00 64-bit"]
[Black "Revenge 1.0 64-bit"]
[ECO "B15"]
[Result "0-1"]

1. e4 {book 0s} c6 {book 0s} 2. Nc3 {book 0s} d5 {book 0s}
3. d4 {book 0s} dxe4 {book 0s} 4. Nxe4 {book 0s} Nf6 {book
0s} 5. Nxf6+ {book 0s} exf6 {book 0s} 6. c3 {book 0s} Bd6
{book 0s} 7. Bd3 {book 0s} O-O {book 0s} 8. Qc2 {book 0s}
h6 {book 0s} 9. Kf1 {-0.20/19 55s} c5 {-0.89/29 36s (Re8)}
10. dxc5 {+0.20/30 24s} Bxc5 {-0.88/30 52s (Bc7)} 11. Bf4
{+0.20/25 1:40m} Qb6 {-0.74/28 1:42m (Kh8)} 12. Rd1
{+0.20/28 18s (Nf3)} a5 {-1.08/27 23s (f5)} 13. b4
{+0.10/23 3:20m (Nf3)} axb4 {-4.00/28 47s (Be7)} 14. cxb4
{-0.20/25 14s} Qxb4 {-4.15/28 59s (Bxb4)} 15. Ne2 {-0.20/25
18s (Bxb8)} Nc6 {-4.45/29 30s (Bg4)} 16. Rb1 {-0.20/26 14s
(Bc4)} Qa3 {-4.30/26 20s} 17. Rb3 {-0.20/29 12s (Kg1)} Qxa2
{-6.04/27 28s} 18. Rb2 {-2.64/28 1:54m (Qxa2)} Qa1+
{-6.59/27 1:11m (Qa7)} 19. Rb1 {-1.76/30 2:00m} Qa7
{-6.31/30 17s (Qa3)} 20. Be4 {-3.76/26 41s (h3)} Re8
{-6.90/27 25s (Bxf2)} 21. Bf3 {-4.95/26 33s (Ng3)} Bxf2
{-7.08/27 36s} 22. h3 {-5.43/27 22s (h4)} Be3 {-7.20/27
17s} 23. Rd1 {-6.04/29 20s} Be6 {-7.61/27 16s (Bxf4)} 0-1[/pgn]

[pgn][Event "40 Moves in 16 min"]
[Site "Revenge Gauntlet, GRAHAM-Z87"]
[Date "2021.07.26"]
[Round "3.3"]
[White "Revenge 1.0 64-bit"]
[Black "Ethereal 13.00 64-bit"]
[ECO "A33"]
[Result "1-0"]

1. c4 {book 0s} c5 {book 0s} 2. Nf3 {book 0s} Nc6 {book 0s}
3. d4 {book 0s} cxd4 {book 0s} 4. Nxd4 {book 0s} Nf6 {book
0s} 5. Nc3 {book 0s} e6 {book 0s} 6. a3 {book 0s} Bc5 {book
0s} 7. Nb3 {book 0s} Bb6 {book 0s} 8. e4 {book 0s} O-O
{book 0s} 9. Be2 {+0.06/26 36s} a5 {+0.20/22 26s (d6)}
10. c5 {+0.43/28 44s} a4 {+0.20/26 38s (Bc7)} 11. cxb6
{+0.60/29 23s} axb3 {+0.20/28 22s} 12. Bf4 {+0.60/31 27s
(Qxb3)} e5 {+0.20/28 22s (Qxb6)} 13. Be3 {+1.14/29 50s} Ne8
{+0.20/29 34s (d6)} 14. Qxb3 {+2.94/29 22s (Bb5)} Nd4
{+0.20/32 28s (d6)} 15. Bxd4 {+3.48/34 28s} exd4 {+0.20/34
1:00m} 16. Nd5 {+3.56/33 1:24m} Nf6 {+0.20/33 17s} 17. O-O
{+3.56/34 25s (Nc7)} Nxd5 {+0.20/30 25s} 18. exd5 {+3.56/33
33s} Qf6 {+0.20/32 28s (Re8)} 19. Rfc1 {+4.29/28 29s (Bd3)}
Qd6 {+0.20/31 25s (Re8)} 20. Rc7 {+5.39/30 55s (Bd3)} g6
{+0.20/32 26s (Qe5)} 21. Rd1 {+6.36/29 1:04m (Bd3)} Kg7
{+0.20/28 1:28m (Qe5)} 22. Rxd4 {+6.73/29 23s (g3)} Qf6
{+3.56/29 1:41m (Re8)} 23. Qb4 {+6.96/26 21s (Rdc4)} d6
{+3.70/31 1:07m} 24. h3 {+7.12/28 36s (Rf4)} h5 {+3.62/27
1:01m} 25. a4 {+7.45/27 25s (Rf4)} Qe5 {+3.29/27 1:15m
(Rb8)} 26. Re4 {+7.71/24 18s} Qxd5 {+4.44/27 21s (Qg5)}
27. Bf3 {+7.81/27 24s} Be6 {+4.94/27 23s (Qg5)} 28. Re3
{+8.22/28 16s} Qa2 {+5.24/29 37s} 29. b3 {+8.09/28 17s
(Ra3)} Qa1+ {+5.83/25 31s (Bxb3)} 30. Re1 {+9.87/27 28s
(Kh2)} 1-0[/pgn]
+20 or -20 eval on every turn. It means the NNUE file is not loaded. Something I fixed and aborted on in 13.25, which is why Ray had issues initially since Cutechess was sending <empty> unexpectedly.

I believe that would imply that the Shredder GUI perhaps requires absolute file paths for the weight files instead of relative ones? Each GUI seems to be different. The advice I've given all those who have E13 is to get the absolute file path from the file explorer. Something like "C:/Users/Andrew/Desktop/MyEngines/..".
Thanks, Andrew. I'll try that.
These current games won't count for rating purposes.

Re: Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems

Posted: Thu Jul 29, 2021 8:38 am
by AlexChess
AndrewGrant wrote: ↑Wed Jul 28, 2021 11:42 pm
CMCanavessi wrote: ↑Wed Jul 28, 2021 10:57 pm
AlexChess wrote: ↑Sat Jul 24, 2021 4:57 pm Hi CMCanavessi,

have you solved all the crashes experienced using Ethereal 13.00?

Regards, AlexChess
Nope. Andrew says there's a problem with Ethereal itself, so I'm waiting for a fixed release.
? We resolved your pondering issue if I recall. https://github.com/AndyGrant/Ethereal/c ... 41943eedc3 and then I sent you updated binaries and you confirmed them to be working? Did I miss a posting from you? If so do you have any more UCI logs from these pondering games? They seem to be the only source of issues ever.

If that did not fix it, I will try again. I just need to know, as I never see the problem myself.

1. Alexchess is trolling
2. Mcclane is trolling him back I think, or maybe not?
1. I'm not trolling! Is not even allowed to test open source engines showing disappoint?
2. Aren't you happy that someone warn you about your engine bugs FOR FREE?

ARM64 will be very popular from November 2021 with the release of Windows 11 ARM64. I have reported you severe issues and you simply ignore them, not supporting your fans. All the other x64 engines run fine on WinARM64, Ethereal 13 standard github crashes using more than 1 CPU (it calculates but never makes a move and loses all games for time. Hash suddenly drops from 80 MB to 0 and stops analysis) Ethereal 12.75 NNUE official (Etherlito) is perfectly stable with 4 CPUs and quite strong, so I will continue to test it trying all SF 20 MB NNUEs to verify the influence of them on an average engine. At infinite level Ethereal 13.xx ssse3 works fine and never stops, so isn't an hardware limitation the cause of my issue. You have also said that would be TRIVIAL for you to compile it for ARM64 knowing the CPU I'm using. I have specified my hardware more times, it's also shown on bottom, but we have never really considered to solve my issue.

3. All other programmers here are very kind and help fans when they need. I'm happy to test their engines and chess GUIs and support them.

Best regards,
AlexChess

Re: Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems

Posted: Thu Jul 29, 2021 5:20 pm
by CMCanavessi
AlexChess wrote: ↑Thu Jul 29, 2021 8:38 am ...
Alex, Ethereal was working ok in my system after the previous update by Andrew (13.07). It was all due to Banksia messing chess 960. I just got a new update I'm testing now (13.25) and all seems well with that one too.

Re: Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems

Posted: Thu Jul 29, 2021 6:47 pm
by amanjpro
CMCanavessi wrote: ↑Thu Jul 29, 2021 5:20 pm
AlexChess wrote: ↑Thu Jul 29, 2021 8:38 am ...
Alex, Ethereal was working ok in my system after the previous update by Andrew (13.07). It was all due to Banksia messing chess 960. I just got a new update I'm testing now (13.25) and all seems well with that one too.
But Alex is running them under ARM x64 I believe, which might not be what you are running Ethereal on?



My two cents to Andrew, calling your users trolls, just because you don't like them (for whatever reason), is very cheap... I hope you appologize for it

Re: Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems

Posted: Thu Jul 29, 2021 7:01 pm
by AndrewGrant
amanjpro wrote: ↑Thu Jul 29, 2021 6:47 pm
CMCanavessi wrote: ↑Thu Jul 29, 2021 5:20 pm
AlexChess wrote: ↑Thu Jul 29, 2021 8:38 am ...
Alex, Ethereal was working ok in my system after the previous update by Andrew (13.07). It was all due to Banksia messing chess 960. I just got a new update I'm testing now (13.25) and all seems well with that one too.
But Alex is running them under ARM x64 I believe, which might not be what you are running Ethereal on?

My two cents to Andrew, calling your users trolls, just because you don't like them (for whatever reason), is very cheap... I hope you appologize for it
Its a troll because I've stated many times that running an x86 through an ARM emulator is not something I can support at this time. I am sorry to see it not work, but it is just the case. Will it change in the future? Probably. I'm in the process of getting access to an AVX512 machine for that support, as well as buying a new smart phone with ARMV8 on it expressly for chess so I can have things local.

Similar to how Ethereal does not run under Wine, although different, since that will not be addressed as source code exists and Wine is not needed.

Re: Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems

Posted: Thu Jul 29, 2021 7:05 pm
by AlexChess
amanjpro wrote: ↑Thu Jul 29, 2021 6:47 pm
CMCanavessi wrote: ↑Thu Jul 29, 2021 5:20 pm
AlexChess wrote: ↑Thu Jul 29, 2021 8:38 am ...
Alex, Ethereal was working ok in my system after the previous update by Andrew (13.07). It was all due to Banksia messing chess 960. I just got a new update I'm testing now (13.25) and all seems well with that one too.
But Alex is running them under ARM x64 I believe, which might not be what you are running Ethereal on?



My two cents to Andrew, calling your users trolls, just because you don't like them (for whatever reason), is very cheap... I hope you appologize for it
Thank you amanjpro, I'm testing in my SuperBlitz 51 (link in the signature) also Zahak 5.0 that works perfectly on Windows 10 ARM64, being an x64 never crashes and it's stronger than previous ...Waiting 6.0 with NNUE, ready as your personal ARM64 betatester, native or not :D

Just started a match Zahak 5.0 vs Ethereal 13 standard github no NNUE both with 1 CPU
Image

PS: Note the Kurdistan flag added to Arena 3.5.1 :wink:

Re: Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems

Posted: Sat Jul 31, 2021 1:13 am
by CMCanavessi
AndrewGrant wrote: ↑Thu Jul 29, 2021 7:01 pm ...
Andrew, I've sent you a pm with logs from a game

Re: Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems

Posted: Sat Jul 31, 2021 2:45 am
by AndrewGrant
CMCanavessi wrote: ↑Sat Jul 31, 2021 1:13 am
AndrewGrant wrote: ↑Thu Jul 29, 2021 7:01 pm ...
Andrew, I've sent you a pm with logs from a game
http://talkchess.com/forum3/viewtopic.p ... e21eba3a44

And often debated topic. I'll change the behavior, now that there is a GUI doing it as stated by UCI instead of as-expected by many.

Re: Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems

Posted: Wed Aug 04, 2021 2:46 pm
by AlexChess
INDECENT REQUEST? :D

Andrew, could you please release a complete Ethereal 13.26 NNUE commercial ssse3 TRIAL (expiring after 1 hour) to test if it works fine with 4 CPUs like Ethereal 12.75 SF NNUE does PERFECTLY on Windows 11 ARM64?

...Only to understand what is happening to my beta Windows 11 ARM64 that works with all other popcount ssse3 | sse41 engines and WinARM64 builds like Igel 3.0.5 (neon) NNUE Arm64 and RubiChess 2.2 WinArm64 (neon) ? :)

Or maybe you could test Ethereal 13 NNUE on a Microsoft Surface Windows 10 | 11 ARM64...
https://www.theverge.com/2019/11/5/2094 ... pecs-price

THANK YOU!
Best regards, AlexChess
(the troll :shock: )
All times are UTC+02:00
Page 27 of 29

Powered by phpBB® Forum Software © phpBB Limited