I really need your help testing Chispa

[Federico Corigliano]

Geschrieben von: / Posted by: Federico Corigliano at 10 January 2004 05:42:44:

I developed Chispa 4.0x sofar without sufficient testing, and is time now to do it. As a single person sometimes isn't sufficient (specially if that person have to do other things at same time), you can download from the Chispa beta's page (link below) a copy of the latest beta and a small openings book.
I need mainly bug reports. Then, when Chispa is clean of bugs, I will need a lot of games to see the rating difference with v3.98d, to measure the strenght lose by the slowdown (about a 10-15%) of this version.
I will really appreciate any e-mail sent to me with a bug report or played games.
Ahh, Geocities have a limit of downloads, so it's possible that you will have to wait an hour to download Chispa.
Federico


http://ar.geocities.com/fedecorigliano/

[Dann Corbit]

Geschrieben von: / Posted by: Dann Corbit at 10 January 2004 06:32:33:
Als Antwort auf: / In reply to: I really need your help testing Chispa geschrieben von: / posted by: Federico Corigliano at 10 January 2004 05:42:44:

I developed Chispa 4.0x sofar without sufficient testing, and is time now to do it. As a single person sometimes isn't sufficient (specially if that person have to do other things at same time), you can download from the Chispa beta's page (link below) a copy of the latest beta and a small openings book.
I need mainly bug reports. Then, when Chispa is clean of bugs, I will need a lot of games to see the rating difference with v3.98d, to measure the strenght lose by the slowdown (about a 10-15%) of this version.
I will really appreciate any e-mail sent to me with a bug report or played games.
Ahh, Geocities have a limit of downloads, so it's possible that you will have to wait an hour to download Chispa.
Federico

While running a test suite, I saw the "proverbial happy faces" in the output. That is usually a sign of something very bad. Here is some sample output that shows the strangeness:
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Partial results: 17 / 28
Cleaning the transposition hash tables... Ready
Cleaning the evaluation hash table... Ready
Position Nº 29:
8/8/8/1p5r/p1p1k1pN/P2pBpP1/1P1K1P2/8 b - - bm Rxh4 b4; id "WAC.229";
1. 280 2 19 ) 1. ... Re5 [nps:950|depth:1/4]
2. 263 5 104 ) 1. ... Re5 2. Kc3 [nps:2080|depth:2/5|BF:5.47]
3 265 10 154 ) 1. ... Re5 2. Kc3 Re6 [nps:1540|depth:3/6]
3. 265 15 321 ) 1. ... Re5 2. Kc3 Re6 [nps:2140|depth:3/7|BF:3.09]
4 265 20 501 ) 1. ... Re5 2. Kc3 Re6 3. Bf4 [nps:2505|depth:4/7]
4. 265 35 828 ) 1. ... Re5 2. Kc3 Re6 3. Bf4 [nps:2365|depth:4/8|BF:2.58]
5 268 36 1313 ) 1. ... Re5 2. Ng6 Re6 3. Nf4 Re5 [nps:3647|depth:5/10]
5. 268 41 2066 ) 1. ... Re5 2. Ng6 Re6 3. Nf4 Re5 [nps:5039|depth:5/10|BF:2.50]
6 252 46 3133 ) 1. ... Re5 2. Ng6 Re8 3. Kc3 Re6 4. Nf4 [nps:6810|depth:6/11]
6 264 52 6415 ) 1. ... Rd5 2. Bf4 Kd4 3. Ng6 Rh5 4. Ne7 [nps:12336|depth:6/12]
6. 264 55 7528 ) 1. ... Rd5 2. Bf4 Kd4 3. Ng6 Rh5 4. Ne7 [nps:13687|depth:6/12|BF:3.64]
7 264 61 11329 ) 1. ... Rd5 2. Bf4 Rh5 3. Be3 '▒↕ 4. ╕╙█♥ 5. ►┐5♥ 6. ►┐5♥ [nps:18572|depth:
7/15]
7. 264 67 19297 ) 1. ... Rd5 2. Bf4 Rh5 3. Be3 ╟╜↕ 4. ╕╙█♥ 5. ►┐5♥ 6. ►┐5♥ [nps:28801|depth:
7/16|BF:2.56]
8 250 79 33407 1. ... Rd5 2. Ng6 Rd8 3. Kc3 Re8 4. Nf4 Re5 5. Bb6 [nps:42287|depth:8/16]
8. 250 92 55350 1. ... Rd5 2. Ng6 Rd8 3. Kc3 Re8 4. Nf4 Re5 5. Bb6 [nps:60163|depth:8/17|BF:2.87]
9 250 97 68209 1. ... Rd5 2. Ng6 Rd8 3. Kc3 Re8 4. Nf4 Re5 5. Bb6 Re8 [nps:70318|depth:9/16]
9. 250 112 109613 1. ... Rd5 2. Ng6 Rd8 3. Kc3 Re8 4. Nf4 Re5 5. Bb6 Re8 [nps:97868|depth:9/20|BF:1.98]
10 250 124 144507 1. ... Rd5 2. Ng6 Rd6 3. Nf4 Rd7 4. Kc3 Re7 5. ╕╙█♥ 6. [nps:116537|depth:10/20]
10 285 172 290531 ) 1. ... Rd5 ++
10 374 244 513739 ) 1. ... b4 2. Ke1 bxa3 3. bxa3 c3 4. Kd1 Rb5 5. Bc1 c2+ 6. Ke1 Rb1 7. Kd2 [nps:210548|depth:10/25]
10. 374 251 530095 ) 1. ... b4 2. Ke1 bxa3 3. bxa3 c3 4. Kd1 Rb5 5. Bc1 c2+ 6. Ke1 Rb1 7. Kd2 [nps:211193|depth:10/25|BF:4.84]
11 409 323 743207 ) 1. ... b4 ++
11 438 422 1067101 ) 1. ... b4 2. axb4 Rxh4 3. Kc3 Rh2 4. b5 Kd5 5. b6 Rh8 6. Kb4 Ra8 [nps:252867|depth:11/28]
Solution FOUND
Partial results: 18 / 29
Cleaning the transposition hash tables... Ready
Cleaning the evaluation hash table... Ready
Position Nº 30:
8/k1b5/P4p2/1Pp2p1p/K1P2P1P/8/3B4/8 w - - bm b6+ Be3; id "WAC.100";
1. 236 2 15 1. Kb3 [nps:750|depth:1/4]
2. 226 4 58 1. Kb3 Kb6 [nps:1450|depth:2/5|BF:3.87]
3 233 9 88 1. Kb3 Kb6 2. Kc3 [nps:977|depth:3/5]
3. 233 14 145 1. Kb3 Kb6 2. Kc3 [nps:1035|depth:3/7|BF:2.50]
4 233 18 199 1. Kb3 Kb6 2. Kc3 Bd6 [nps:1105|depth:4/7]
4. 233 23 394 1. Kb3 Kb6 2. Kc3 Bd6 [nps:1713|depth:4/9|BF:2.72]
5 241 29 940 1. Kb3 Bd6 2. Kc2 Kb6 3. Kd3 [nps:3241|depth:5/11]
5. 241 34 1228 1. Kb3 Bd6 2. Kc2 Kb6 3. Kd3 [nps:3611|depth:5/11|BF:3.12]
6 241 39 1848 1. Kb3 Bd6 2. Kc2 Kb6 3. Kd3 Bc7 [nps:4738|depth:6/13]
6. 241 44 2426 1. Kb3 Bd6 2. Kc2 Kb6 3. Kd3 Bc7 [nps:5513|depth:6/13|BF:1.98]
7 239 50 5523 1. Kb3 Bb8 2. Be3 Kb6 3. Kc3 Bd6 4. Kd3 [nps:11046|depth:7/15]
7. 239 55 6942 1. Kb3 Bb8 2. Be3 Kb6 3. Kc3 Bd6 4. Kd3 [nps:12621|depth:7/16|BF:2.86]
8 239 61 11290 1. Kb3 Bb8 2. Be3 Kb6 3. Kc3 Bd6 4. Kd3 Bc7 [nps:18508|depth:8/20]
8. 239 65 14285 1. Kb3 Bb8 2. Be3 Kb6 3. Kc3 Bd6 4. Kd3 Bc7 [nps:21976|depth:8/23|BF:2.06]
9 241 72 23766 1. Kb3 Bd6 2. Be3 Kb6 3. Kc3 Bb8 4. Kd3 Bd6 5. Bd2 [nps:33008|depth:9/21]
9. 241 76 29605 1. Kb3 Bd6 2. Be3 Kb6 3. Kc3 Bb8 4. Kd3 Bd6 5. Bd2 [nps:38953|depth:9/22|BF:2.07]
10 241 87 37837 1. Kb3 Bd6 2. Be3 Kb6 3. Kc3 Bb8 4. Kd3 Bd6 5. Bd2 Bc7 [nps:43490|depth:10/21]
10. 241 91 49929 1. Kb3 Bd6 2. Be3 Kb6 3. Kc3 Bb8 4. Kd3 Bd6 5. Bd2 Bc7 [nps:54867|depth:10/26|BF:1.69]
11 241 97 63683 1. Kb3 Bd6 2. Be3 Kb6 3. Kc3 Bb8 4. Kd3 Bd6 5. Kc3 '▒↕ 6. ►┐5♥ 7. ►┐5♥ 8. °B☻♦ [nps:65652|depth:11/26]
11. 241 108 85331 1. Kb3 Bd6 2. Be3 Kb6 3. Kc3 Bb8 4. Kd3 Bd6 5. Kc3 ╟╜↕ 6. ►┐5♥ 7. ►┐5♥ 8. °B☻♦ [nps:79010|depth:11/26|BF:1.71]
12 253 166 283478 ) 1. Kb3 Bd6 2. Kc3 Bc7 3. Kd3 Bd6 4. Ke3 Be7 5. Bc3 Kb6 6. Bb2 Kc7 [nps:170769|depth:12/28]
12. 253 179 330321 ) 1. Kb3 Bd6 2. Kc3 Bc7 3. Kd3 Bd6 4. Ke3 Be7 5. Bc3 Kb6 6. Bb2 Kc7 [nps:184536|depth:12/30|BF:3.87]
13 256 232 524541 ) 1. Kb3 Bd6 2. Kc3 Bc7 3. Kd3 Bb8 4. Ke3 Bd6 5. Bc3 Be7 6. ª▒@ 7. [nps:226095|depth:13/32]
13. 256 262 634290 ) 1. Kb3 Bd6 2. Kc3 Bc7 3. Kd3 Bb8 4. Ke3 Bd6 5. Bc3 Be7 6. ª▒@ 7. [nps:242095|depth:13/33|BF:1.92]
14 264 321 835968 ) 1. Kb3 Bd6 2. Kc3 Bc7 3. Kd3 Bb8 4. Ke3 Bd6 5. Ba5 Bf8 6. Bd8 Bg7 7. Bc7 Ka8 [nps:260426|depth:14/32]
14. 264 390 1099206 ) 1. Kb3 Bd6 2. Kc3 Bc7 3. Kd3 Bb8 4. Ke3 Bd6 5. Ba5 Bf8 6. Bd8 Bg7 7. Bc7 Ka8 [nps:281847|depth:14/38|BF:1.73]
Solution NOT FOUND
Partial results: 18 / 30
Cleaning the transposition hash tables... Ready
Cleaning the evaluation hash table... Ready
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Here is the start:
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
E:\programme\winboard\chispa>chispa_4x.exe
*----------------------------------------------
| Chispa 4x 01-10 - WinBoard and UCI Chess Engine
| Author: Federico Andrés Corigliano
| Buenos Aires - Argentina
| Compilation date: Jan 9 2004 - 17:00:27
| Contact me: fedecori@hotmail.com
*----------------------------------------------
Deallocating the transpositions hash tables memory... Ready
Allocating memory for the transposiciones hash tables...
Allocated 2097152 hash entries = 33554432 bytes
Cleaning the transposition hash tables... Ready
Deallocating the evaluation hash table memory... Ready
Allocating memory for the evaluation hash table...
Allocated 262144 ehash entries = 4194304 bytes
Cleaning the evaluation hash table... Ready
Deallocating the pawns hash table memory... Ready
Allocating memory for the pawns hash table...
Allocated 65536 phash entries = 3670016 bytes
Cleaning the pawns hash table... Ready
Checking for endgame tablebases in e:\programme\winboard\nalimov
Using 5-men endgame tablebases
Allocated 8388608 bytes for EGTB cache
Chispa 4x 01-10> help
avaimoves : Shows the available moves for the position
ayuda : Shows this screen
board : Shows the board
book : Starts the Openings Book Edition mode
bookoff : Sets the use of the book off
bookon : Sets the use of the book on
divide : Same as Perft but divided by the root moves
eval : Shows the static eval for the current position
eval-epd-test : Test the evaluation routine with an EPD file
fen : Set a FEN position
hash : Set the number of MBs for the hash tables
hashkey : Shows the hash key for the current position
help : Shows this screen
history : Shows a list of played moves
idioma : Changes the language of Chispa
language : Changes the language of Chispa
libro : Starts the Openings Book Edition mode
log : Enable or disable the logging of commands
mirror : Reflects the board and the color
new : Start a new game
nuevo : Start a new game
perf : Show the move generation performance
perft : Calculate the total nodes until a desired depth
quit : Exit from Chispa
Press ENTER to continue...
salir : Exit from Chispa
tablero : Shows the board
testepd : Run an EPD test-suite
variant : To choose a chess variant
uci : Starts UCI mode
xboard : Starts Winboard mode
Chispa 4x 01-10> testepd
EPD Filename: \epd\tprobs.epd
Seconds per position: 5
Cleaning the transposition hash tables... Ready
Cleaning the evaluation hash table... Ready
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Here is the end:
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Solution FOUND
Partial results: 34 / 48
Nodes: 62019495 Qnodes: 17586538 (28.356%) Ordering efficiency: 94.578%
Transposition Hash: (S/P/H): 15427111/19417717/8490115 (43.724%)
Pawn Hash: (S/P/H): 394087/8457563/8063476 (95.340%)
Evaluation Hash: (S/P/H): 39840016/54761319/14921303 (27.248%)
Egtb: (P/H): 1/1 (50.000%)
Null-Move: (T/C): 10487389/8442398 (80.500%)
Extensions: (C/OR/RC/T/PP/EC): 2416910/227710/633737/69811/329909/53478
Evaluations: (F/L): 8457563/31382453
Re-searchs: 94 IID: 195 See cuts: 14940181
FINAL RESULTS:
---------------
Found positions -> 34 (70.833%)
Not found positions -> 14 (29.167%)
-----------------------------
Total positions -> 48 (100.00%%)
Total time -> 249.570 seg.
Chispa 4x 01-10>
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Here is my ini file:
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
E:\programme\winboard\chispa>type chispa.ini
[Hash Tables]
HashMB: 32
PHashMB: 4
EHashMB: 4
[Endgame Tablebases]
EGTBDir: e:\programme\winboard\nalimov
EGTBCacheMB: 8
[Openings book]
OpeningsBook: TRUE
OPNBookFile: libro.bin
[Resign]
Resign: TRUE
ResignValue: -650
[Search]
AspirationWindow: 35
[NullMove]
NullMove: TRUE
NullMoveMax: 300
NullMoveMin: 200
[Extensions]
CheckExtension: 100
RecapExtension: 75
SingleRepCheckExtension: 75
MateThreatExtension: 75
PawnPushExtension: 75
[Evaluation]
PawnValue: 100
KnightValue: 315
BishopValue: 335
RookValue: 550
QueenValue: 1050
LazyEval: TRUE
[Logging]
CreateLog: TRUE
LogInSingleFile: FALSE
[Pruning]
Futility: FALSE
AlphaPrune: FALSE
Reduc1: FALSE
[Other]
ExtraInfo: TRUE
Language: ENG
Variant: NONE
; Explanations:
; Pieces values and aspiration window values are in centipawns (PAWN = 100)
; Search extensions values and nullmove values are in centiplies (FULLPLY = 100)
; Language possible values: ENG ->English
; ESP ->Spanish
; Variant possible values: FRC ->Fischer Random Chess
; Other thing at your choice ->Normal Chess



my ftp site {remove http:// unless you like error messages}

[Patrick Buchmann]

Geschrieben von: / Posted by: Patrick Buchmann at 10 January 2004 19:58:55:
Als Antwort auf: / In reply to: I really need your help testing Chispa geschrieben von: / posted by: Federico Corigliano at 10 January 2004 05:42:44:

I developed Chispa 4.0x sofar without sufficient testing, and is time now to do it. As a single person sometimes isn't sufficient (specially if that person have to do other things at same time), you can download from the Chispa beta's page (link below) a copy of the latest beta and a small openings book.
I need mainly bug reports. Then, when Chispa is clean of bugs, I will need a lot of games to see the rating difference with v3.98d, to measure the strenght lose by the slowdown (about a 10-15%) of this version.
I will really appreciate any e-mail sent to me with a bug report or played games.
Ahh, Geocities have a limit of downloads, so it's possible that you will have to wait an hour to download Chispa.
Federico

Hi Federico,
A "tournoi d'entrée" in the UEL is running (80 games).
Chispa 4x 01-10 run under Arena and Fritz GUI, but not under Shredder Classic.
Patrick


Le Fou numérique

[Patrick Buchmann]

Geschrieben von: / Posted by: Patrick Buchmann at 10 January 2004 22:58:09:
Als Antwort auf: / In reply to: Re: I really need your help testing Chispa geschrieben von: / posted by: Patrick Buchmann at 10 January 2004 19:58:55:

Hi Federico,
A "tournoi d'entrée" in the UEL is running (80 games).
Chispa 4x 01-10 run under Arena and Fritz GUI, but not under Shredder Classic.
Patrick

Hi Federico,
The tournament is finish.
The rating UEL is 2370 just before 3.98d.
Program Ruel + - Games Score Av.Op. Draws
1 Ruffian 1.0.5 : 2661 31 50 257 72.8 % 2490 24.1 %
2 SOS 4 for Arena : 2642 42 82 132 76.9 % 2434 18.9 %
3 Aristarch 4.21 : 2613 26 30 465 63.0 % 2520 27.5 %
4 Yace Paderborn : 2578 27 26 525 57.1 % 2528 24.0 %
5 Pepito v1.59 : 2549 28 25 533 53.6 % 2525 24.6 %
6 Ktulu 4.2 : 2548 56 100 80 76.2 % 2345 20.0 %
7 Pharaon 2.62 : 2535 24 23 706 55.1 % 2499 22.4 %
8 AnMon 5.21 : 2532 23 22 776 55.1 % 2497 20.5 %
9 Little Goliath Nemesis : 2530 34 28 419 50.7 % 2525 21.7 %
10 Tao 5.4 : 2527 24 22 750 54.3 % 2497 19.6 %
11 Abrok 5.0 : 2490 27 34 411 49.5 % 2494 26.0 %
12 Gothmog 0.4.3 : 2481 50 60 138 62.0 % 2396 18.1 %
13 Amyan 1.592 : 2479 53 47 169 51.5 % 2469 16.6 %
14 Dragon 4.5 CF : 2476 47 50 168 59.5 % 2409 21.4 %
15 Patzer 3.61 : 2474 38 68 176 72.2 % 2309 15.9 %
16 Nejmet 3.07 : 2474 27 33 405 47.8 % 2489 26.4 %
17 Sjeng 12.13 : 2473 27 25 587 54.1 % 2444 19.9 %
18 Frenzee 146 : 2452 38 41 269 57.2 % 2401 15.6 %
19 Queen 2.42a : 2443 52 50 150 55.7 % 2403 20.7 %
20 Terra 3.2 : 2399 47 50 184 48.4 % 2410 12.0 %
21 Leila 0.42u : 2390 80 83 65 56.9 % 2342 15.4 %
22 LambChop 10.99 : 2388 72 63 102 43.6 % 2432 6.9 %
23 Chispa 4x 01-10 : 2370 77 77 80 51.9 % 2357 6.2 %
24 Chispa v3.98d : 2342 54 61 130 49.6 % 2344 14.6 %
25 Hagrid 0.7.56 : 2340 29 31 477 48.2 % 2352 10.9 %
26 Madeleine 0.2 : 2327 33 31 420 51.8 % 2315 12.6 %
27 Hermann 0.3.2 : 2285 43 36 286 41.4 % 2345 10.1 %
28 DelphiMAX 2.8 : 2275 58 60 134 53.4 % 2251 6.7 %
29 Monarch 2002-04c : 2250 61 45 165 38.2 % 2334 9.7 %
30 Taktix 2.23f : 2236 54 50 166 50.6 % 2232 9.6 %
31 BigLion 2.23f : 2226 73 46 140 33.2 % 2347 9.3 %
32 Asterisk 0.4b : 2186 104 59 80 27.5 % 2354 12.5 %
33 Celes v0.77c : 2154 39 40 276 46.6 % 2178 12.7 %
34 EnginMax 5.11c : 2154 44 40 249 44.0 % 2196 10.0 %
35 WJChess v1.52 : 2146 94 55 94 29.3 % 2299 11.7 %
36 Silke Chess 1.2.1209 : 2126 45 33 317 38.5 % 2207 8.2 %
37 Simontacchi v1.8a : 2083 64 39 198 34.3 % 2195 5.1 %
38 Gaia 1.1 Massy : 2063 61 42 175 35.4 % 2167 11.4 %
39 KKFChess v2.6.0pmx : 2008 193 50 78 10.9 % 2373 6.4 %
40 Piranha 0.5 : 1976 80 29 262 21.8 % 2198 4.6 %
41 FUSC# V1.09 : 1767 347 0 72 2.8 % 2367 2.8 %
42 T.rex 1.8.5 : 1733 208 27 220 5.7 % 2221 2.3 %
Tournament table and PGN with 80 games on my homepage.
Patrick


Le Fou numérique - Matchs et tournois entre Engines

[Federico Corigliano]

Geschrieben von: / Posted by: Federico Corigliano at 11 January 2004 00:30:34:
Als Antwort auf: / In reply to: Re: I really need your help testing Chispa geschrieben von: / posted by: Dann Corbit at 10 January 2004 06:32:33:

The smileys before the PV are OK. The bad are the strange characters in middle of the PV. Seems to be a bug in function CollectPVfromHash(). I will try to find it.
Thanks,
Federico

[Federico Corigliano]

Geschrieben von: / Posted by: Federico Corigliano at 11 January 2004 00:35:00:
Als Antwort auf: / In reply to: Chispa 4x 01-10 : 80 games geschrieben von: / posted by: Patrick Buchmann at 10 January 2004 22:58:09:

Thank you very much Patrick!
The result is a surprise for me, I expected about -25 ELO than v3.98d.
Federico

[Federico Corigliano]

Geschrieben von: / Posted by: Federico Corigliano at 11 January 2004 03:42:07:
Als Antwort auf: / In reply to: Re: I really need your help testing Chispa geschrieben von: / posted by: Federico Corigliano at 11 January 2004 00:30:34:

I found the bug. The problem was a char[] that don't was set to 0 before to put in the screen.
Federico

The smileys before the PV are OK. The bad are the strange characters in middle of the PV. Seems to be a bug in function CollectPVfromHash(). I will try to find it.
Thanks,
Federico