There is a very strange issue with the server at the moment. The only thing that has changed is mamer, In the mamer config header I increased the max. number of rounds from 10 to 12 and recompiled. We also run a test tourney with 20 players and that went fine.
My guess is that the old mamer is compiled with a very old version of GCC, I used a much newer version of GCC to compile mamer and this probably breaks something. The Lasker/Capablanca server software is very sensitive to strange inputs, I think the recompiled mamer sends something to the server that it does not like, or maybe it has an issue with the increased max. number of rounds from 10 to 12, but that should not crash the server.
On-line engine blitz tourney July
Moderators: hgm, Rebel, chrisw
-
- Posts: 1563
- Joined: Thu Jul 16, 2009 10:47 am
- Location: Almere, The Netherlands
-
- Posts: 1563
- Joined: Thu Jul 16, 2009 10:47 am
- Location: Almere, The Netherlands
Re: On-line engine blitz tourney July
After replacing mamer with the old one everything was stable again. Because we lost half an hour due to all the issues we decided to play nine rounds.
In the end Belfish ended on top.
Nightmare was very lucky this time and had a chance to win the tourney, unfortunately he (or it) dropped a half point with a book repetition against Goldbar. I played with exactly the same engine as last month with only a very small change in lazy-smp which somehow makes that it looks deeper than before.
There are 5 engines on a shared third place, the level at the top seems to be very close.
In the end Belfish ended on top.
Nightmare was very lucky this time and had a chance to win the tourney, unfortunately he (or it) dropped a half point with a book repetition against Goldbar. I played with exactly the same engine as last month with only a very small change in lazy-smp which somehow makes that it looks deeper than before.
There are 5 engines on a shared third place, the level at the top seems to be very close.
Code: Select all
:Tourney Players: Round 9 of 9
:
: Name Rating Score Perfrm Upset Results
: ----------------- ------ ----- ------ ------ -------
: 1 +BelFish [2959] 7.0 [3036] [ 59] =14w +07b =02w +12b +09w =03b =04w +05b +06w
: 2 +NightmareX [2844] 6.5 [2975] [ 432] +26w =03b =01b +18w +08w =04b +05w =09b =11w
: 3 +LczTinker [3055] 5.5 [2979] [ 0] +18w =02w =06b +17w =05b =01w +11b =04w *08b
: 4 +Honey [2982] 5.5 [2955] [ 36] =07w =06b +14w +11w +13b =02w =01b =03b *09w
: 5 +Schooner [2926] 5.5 [2832] [ 68] +20w +17w =13b =08b =03w +07w -02b -01w +15b
: 6 +WaspX [2846] 5.5 [2871] [ 288] +22b =04w =03w -10b +14b +08w =09w +17b -01b
: 7 +atomNightmare [2600] 5.5 [2720] [ 344] =04b -01w +23b +19w +10w -05b -08b +18w +16w
: 8 +ArasanX [2934] 5.0 [2839] [ 0] +19w +10w =09b =05w -02b -06b +07w +12b *03w
: 9 +LeelaRB [2902] 5.0 [2881] [ 16] +15w +11b =08w =13w -01b +17w =06b =02w *04b
: 10 +Marvin [2749] 5.0 [2620] [ 97] +16w -08b =19w +06w -07b =12b =15w =11b +18b
: 11 +Goldbar [2736] 5.0 [2712] [ 189] +24b -09w +21w -04b +15w +13b -03w =10w =02b
: 12 +WaDuuttie [2635] 5.0 [2512] [ 57] +25b -13b +22w -01w =18b =10w +21b -08w +24b
: 13 -Nemorino [2865] 4.5 [2615] [ 48] +21b +12w =05w =09b -04w -11w -18b =19b +23w
: 14 +atomGoldbar [2541] 4.5 [2456] [ 240] =01b =18w -04b +22b -06w =16w -17b +26w +25b
: 15 +rpiArminius [2329] 4.5 [2518] [ 249] -09b +24w -17b +16w -11b +22w =10b +20w -05w
: 16 +RavenEngine [2080] 4.5 [2287] [ 452] -10b -19b +25w -15b +24w =14b +23w +21w -07b
: 17 +Amoeba [2739] 4.0 [2617] [ 0] +23w -05b +15w -03b +21w -09b +14w -06w *19b
: 18 +Arminius [2604] 4.0 [2630] [ 276] -03b =14b +20w -02b =12w +19b +13w -07b -10w
: 19 +CoronaVirusChess [2416] 4.0 [2535] [ 390] -08b +16w =10b -07b +20w -18w +26b =13w *17w
: 20 +Myrddin [2368] 4.0 [2234] [ 0] -05b =23w -18b +26w -19b =21w +25b -15b +22w
: 21 +CTdccc [2302] 3.5 [2257] [ 33] -13w +25w -11b +24w -17b =20b -12w -16b +26b
: 22 +feligres [2186] 3.5 [2183] [ 0] -06w +26b -12b -14w +23b -15b =24w +25w -20b
: 23 +Abbess [2038] 3.0 [2148] [ 211] -17b =20b -07w +25b -22w +26w -16b =24w -13b
: 24 +Skiull [1982] 3.0 [2091] [ 232] -11w -15b +26w -21b -16b +25w =22b =23b -12w
: 25 -Skipper [1629] 1.0 [1935] [ 455] -12w -21b -16b -23w +26b -24b -20w -22b -14w
: 26 +atomSargonI [2084] 0.0 [1856] [ 0] -02b -22w -24b -20b -25w -23b -19w -14b -21w
:
: Average Rating 2551.2
-
- Posts: 3547
- Joined: Thu Jun 07, 2012 11:02 pm
Re: On-line engine blitz tourney July
LeelaRB was playing on a low cost Ryzen 7 3700U laptop, using the DX12 version of v 0.26.0 running on the integrated Vega 10 graphics. It lost one game horribly, not sure why that was, but otherwise it seemed to play pretty well.
-
- Posts: 27796
- Joined: Fri Mar 10, 2006 10:06 am
- Location: Amsterdam
- Full name: H G Muller
Re: On-line engine blitz tourney July
Indeed, for the tourney type you can also choose 'm' for McMahon, next to 's' for Swiss and 'r' for round robin.Joost Buijs wrote: ↑Sat Jul 11, 2020 3:52 pmI assume that all mamer commands remain the same, and that it is just a small modification to the Swiss format?
-
- Posts: 1563
- Joined: Thu Jul 16, 2009 10:47 am
- Location: Almere, The Netherlands
Re: On-line engine blitz tourney July
Before the tournament Nightmare played some games against LeelaRB and I noticed that a few times Leela came out of book in a (tactically) lost position, maybe this happened at that one game too.Modern Times wrote: ↑Sat Jul 11, 2020 11:09 pm LeelaRB was playing on a low cost Ryzen 7 3700U laptop, using the DX12 version of v 0.26.0 running on the integrated Vega 10 graphics. It lost one game horribly, not sure why that was, but otherwise it seemed to play pretty well.
-
- Posts: 1563
- Joined: Thu Jul 16, 2009 10:47 am
- Location: Almere, The Netherlands
Re: On-line engine blitz tourney July
I will discuss it with Liesebeth.hgm wrote: ↑Sat Jul 11, 2020 11:39 pmIndeed, for the tourney type you can also choose 'm' for McMahon, next to 's' for Swiss and 'r' for round robin.Joost Buijs wrote: ↑Sat Jul 11, 2020 3:52 pmI assume that all mamer commands remain the same, and that it is just a small modification to the Swiss format?
First I have to find out what went wrong after I recompiled mamer with the max. number of rounds increased to 12. What I can think of is, that it is possible that it was compiled for 64 bit and that this breaks something, I have to look at the make file to check this. On the other hand it is very sloppy that a remote program can bring the server on it's knees.
I rebuild mamer with:
autoreconv -f -i
./configure "$@"
make clean
make -f makefile -j2
It compiled without a hitch. Liesebeth played a test tourney with 20 players and 11 rounds, that went fine too, so I assumed it all worked.
I noticed that the server (chessd) is very susceptible to malformed input, this is something that needs to be solved too.
-
- Posts: 3547
- Joined: Thu Jun 07, 2012 11:02 pm
Re: On-line engine blitz tourney July
Yes I think that was indeed the case.Joost Buijs wrote: ↑Sun Jul 12, 2020 7:33 amBefore the tournament Nightmare played some games against LeelaRB and I noticed that a few times Leela came out of book in a (tactically) lost position, maybe this happened at that one game too.Modern Times wrote: ↑Sat Jul 11, 2020 11:09 pm LeelaRB was playing on a low cost Ryzen 7 3700U laptop, using the DX12 version of v 0.26.0 running on the integrated Vega 10 graphics. It lost one game horribly, not sure why that was, but otherwise it seemed to play pretty well.
-
- Posts: 4317
- Joined: Tue Apr 03, 2012 4:28 pm
Re: On-line engine blitz tourney July
Fun tournament, mine was with no opening book, no threads code and came in the bottom half on its first outing, but I quite liked this game (playing white). Not sure how sound was 12 Nxd4, the game flashes by too fast to analyse in real time, but I figured the exchange sac was on the cards at that point. Stockfish prefers Rxd4
[pgn] 1. e4 (0:00) c5 (0:00)
2. Nf3 (0:09) d6 (0:00)
3. Nc3 (0:09) Nf6 (0:00)
4. d4 (0:08) cxd4 (0:07)
5. Qxd4 (0:08) e5 (0:07)
6. Qe3 (0:08) Be7 (0:07)
7. Bc4 (0:08) O-O (0:06)
8. O-O (0:08) Bg4 (0:06)
9. h3 (0:07) Bh5 (0:06)
10. b3 (0:07) Nc6 (0:06)
11. Rd1 (0:07) Nd4 (0:06)
12. Nxd4 (0:07) Bxd1 (0:06)
13. Nf5 (0:07) Bh5 (0:06)
14. Nd5 (0:06) Nxd5 (0:06)
15. Bxd5 (0:06) Qc7 (0:05)
16. Qg3 (0:06) Bg6 (0:05)
17. c4 (0:06) Bxf5 (0:05)
18. exf5 (0:06) Bf6 (0:05)
19. Bh6 (0:06) Kh8 (0:05)
20. Bd2 (0:05) Rad8 (0:05)
21. Rd1 (0:05) Rd7 (0:05)
22. Qf3 (0:05) Rfd8 (0:05)
23. Qh5 (0:05) Re8 (0:05)
24. h4 (0:05) Rde7 (0:05)
25. Bg5 (0:05) Bxg5 (0:04)
26. hxg5 (0:05) g6 (0:06)
27. Qh6 (0:04) gxf5 (0:04)
28. Qf6+ (0:04) Kg8 (0:04)
29. g6 (0:04) Re6 (0:04)
30. gxh7+ (0:04) Kxh7 (0:04)
31. Qxf5+ (0:04) Kg7 (0:04)
32. Rd3 (0:04) Rh6 (0:03)
33. Rg3+ (0:04) Kf8 (0:03)
34. Rf3 (0:04) Re7 (0:03)
35. Qg5 (0:04) Rh8 (0:06)
36. Qf6 (0:04) Rg8 (0:03)
37. b4 (0:03) Ke8 (0:03)
38. b5 (0:03) Rf8 (0:03)
39. g3 (0:03) Qc5 (0:03)
40. a4 (0:03) Qb4 (0:03)
41. Kg2 (0:03) Qc5 (0:03)
42. a5 (0:03) Qc7 (0:03)
43. Ra3 (0:03) Rd7 (0:03)
44. a6 (0:03) Qd8 (0:02)
45. Qh6 (0:03) b6 (0:02)
46. Bc6 (0:03) Qe7 (0:26)
47. c5 (0:03) bxc5 (0:02)
48. b6 (0:03) Kd8 (0:04)
49. Rb3 (0:03) c4 (0:05)
50. Rb1 (0:03) axb6 (0:02)
51. Bxd7 (0:03) Kc7 (0:10)
52. Bf5 (0:02) Ra8 (0:02)
53. Rb4 (0:02) b5 (0:04)
54. Rxb5 (0:02) Rxa6 (0:03)
55. Qh8 (0:02) Kc6 (0:02)
56. Rb2 (0:02) Qb7 (0:03)
57. Qe8+ (0:02) Kd5 (0:18)
58. Rxb7 (0:00) Kd4 (0:02)
59. Qb5 (0:00) e4 (0:01)
60. Qxa6 (0:00) Ke5 (0:01)
61. Rxf7 (0:00) Kd4 (0:01)
62. Qxd6+ (0:00) Kc3 (0:01)
63. Bxe4 (0:00) Kb3 (0:01)
64. Rb7+ (0:00) Kc3 (0:01)
65. Qe5+ (0:00) Kd2 (0:00)
66. Qf4+ (0:00) Ke2 (0:01)
67. Qe3+ (0:00) Kd1 (0:00)
68. Rb1+ (0:00)
{Black checkmated} 1-0[/pgn]
[pgn] 1. e4 (0:00) c5 (0:00)
2. Nf3 (0:09) d6 (0:00)
3. Nc3 (0:09) Nf6 (0:00)
4. d4 (0:08) cxd4 (0:07)
5. Qxd4 (0:08) e5 (0:07)
6. Qe3 (0:08) Be7 (0:07)
7. Bc4 (0:08) O-O (0:06)
8. O-O (0:08) Bg4 (0:06)
9. h3 (0:07) Bh5 (0:06)
10. b3 (0:07) Nc6 (0:06)
11. Rd1 (0:07) Nd4 (0:06)
12. Nxd4 (0:07) Bxd1 (0:06)
13. Nf5 (0:07) Bh5 (0:06)
14. Nd5 (0:06) Nxd5 (0:06)
15. Bxd5 (0:06) Qc7 (0:05)
16. Qg3 (0:06) Bg6 (0:05)
17. c4 (0:06) Bxf5 (0:05)
18. exf5 (0:06) Bf6 (0:05)
19. Bh6 (0:06) Kh8 (0:05)
20. Bd2 (0:05) Rad8 (0:05)
21. Rd1 (0:05) Rd7 (0:05)
22. Qf3 (0:05) Rfd8 (0:05)
23. Qh5 (0:05) Re8 (0:05)
24. h4 (0:05) Rde7 (0:05)
25. Bg5 (0:05) Bxg5 (0:04)
26. hxg5 (0:05) g6 (0:06)
27. Qh6 (0:04) gxf5 (0:04)
28. Qf6+ (0:04) Kg8 (0:04)
29. g6 (0:04) Re6 (0:04)
30. gxh7+ (0:04) Kxh7 (0:04)
31. Qxf5+ (0:04) Kg7 (0:04)
32. Rd3 (0:04) Rh6 (0:03)
33. Rg3+ (0:04) Kf8 (0:03)
34. Rf3 (0:04) Re7 (0:03)
35. Qg5 (0:04) Rh8 (0:06)
36. Qf6 (0:04) Rg8 (0:03)
37. b4 (0:03) Ke8 (0:03)
38. b5 (0:03) Rf8 (0:03)
39. g3 (0:03) Qc5 (0:03)
40. a4 (0:03) Qb4 (0:03)
41. Kg2 (0:03) Qc5 (0:03)
42. a5 (0:03) Qc7 (0:03)
43. Ra3 (0:03) Rd7 (0:03)
44. a6 (0:03) Qd8 (0:02)
45. Qh6 (0:03) b6 (0:02)
46. Bc6 (0:03) Qe7 (0:26)
47. c5 (0:03) bxc5 (0:02)
48. b6 (0:03) Kd8 (0:04)
49. Rb3 (0:03) c4 (0:05)
50. Rb1 (0:03) axb6 (0:02)
51. Bxd7 (0:03) Kc7 (0:10)
52. Bf5 (0:02) Ra8 (0:02)
53. Rb4 (0:02) b5 (0:04)
54. Rxb5 (0:02) Rxa6 (0:03)
55. Qh8 (0:02) Kc6 (0:02)
56. Rb2 (0:02) Qb7 (0:03)
57. Qe8+ (0:02) Kd5 (0:18)
58. Rxb7 (0:00) Kd4 (0:02)
59. Qb5 (0:00) e4 (0:01)
60. Qxa6 (0:00) Ke5 (0:01)
61. Rxf7 (0:00) Kd4 (0:01)
62. Qxd6+ (0:00) Kc3 (0:01)
63. Bxe4 (0:00) Kb3 (0:01)
64. Rb7+ (0:00) Kc3 (0:01)
65. Qe5+ (0:00) Kd2 (0:00)
66. Qf4+ (0:00) Ke2 (0:01)
67. Qe3+ (0:00) Kd1 (0:00)
68. Rb1+ (0:00)
{Black checkmated} 1-0[/pgn]
-
- Posts: 1357
- Joined: Wed Mar 08, 2006 10:15 pm
- Location: San Francisco, California
Re: On-line engine blitz tourney July
That looks like the game against Myrddin. Myrddin's horrifically bad king safety code strikes again....chrisw wrote: ↑Sun Jul 12, 2020 10:02 pm Fun tournament, mine was with no opening book, no threads code and came in the bottom half on its first outing, but I quite liked this game (playing white). Not sure how sound was 12 Nxd4, the game flashes by too fast to analyse in real time, but I figured the exchange sac was on the cards at that point. Stockfish prefers Rxd4
-
- Posts: 6991
- Joined: Thu Aug 18, 2011 12:04 pm
Re: On-line engine blitz tourney July
Seems Chris is working on the long awaited Chess System Tal III
90% of coding is debugging, the other 10% is writing bugs.