Perft(13) [3.4 GHz Core i7-2600, 16 GB RAM]

[sje]

Draft 11 record number 68:
[d]rnbqkbnr/pppp1ppp/4p3/8/8/7N/PPPPPPPP/RNBQKB1R w KQkq - 0 2[/d]
The perft(11) for the above is: 7,293,158,888,151,573

68 (17%) down, 332 (83%) to go.

[sje]

Draft 11 record number 69:
[d]rnbqkbnr/pppp1ppp/4p3/8/8/1P6/P1PPPPPP/RNBQKBNR w KQkq - 0 2[/d]
The perft(11) for the above is: 8,211,846,544,768,705

69 (17.25%) down, 331 (82.75%) to go.

[sje]

After 18 weeks, the run has produced 311,484 draft 8 records, about 38% of the total needed.

Mean time per draft 8 record: ca. 35 seconds

Most recent output:

Code: Select all

r1bqkbnr/pppp1ppp/n3p3/8/7N/8/PPPPPPPP/RNBQKB1R w KQkq - 2 3 9 8473017379255
rnbqkb1r/pppp1ppp/5n2/4p3/8/PP6/R1PPPPPP/1NBQKBNR b Kkq - 0 3 8 236028081765
rnb1kbnr/pppp1ppp/8/3Np1q1/8/8/PPPPPPPP/1RBQKBNR b Kkq - 3 3 8 852074768610
rnb1kbnr/pppp1ppp/8/4p1q1/8/2N5/PPPPPPPP/1RBQKBNR w Kkq - 2 3 9 17941288468685
rnb1kbnr/ppppqppp/4p3/6P1/P7/8/1PPPPP1P/RNBQKBNR b KQkq - 0 3 8 295631550395
rnb1kbnr/ppppqppp/4p3/8/P5P1/8/1PPPPP1P/RNBQKBNR w KQkq - 1 3 9 9455269190033
rnbqk1nr/pppp1ppp/4p3/1P6/8/b2P4/P1P1PPPP/RNBQKBNR b KQkq - 0 3 8 534291450411
rnbqkb1r/pppp1ppp/5n2/4p3/1P6/P7/R1PPPPPP/1NBQKBNR b Kkq - 0 3 8 219291357654
rnbqkbnr/pppp1pp1/8/4p2p/8/N2P1N2/PPP1PPPP/R1BQKB1R b KQkq - 0 3 8 782271006818
rnb1kbnr/pppp1ppp/8/4p1q1/5N2/N7/PPPPPPPP/R1BQKB1R b KQkq - 3 3 8 846246982982

[sje]

Draft 11 record number 70:
[d]rnbqkbnr/pppp1ppp/4p3/8/1P6/8/P1PPPPPP/RNBQKBNR w KQkq - 0 2[/d]
The perft(11) for the above is: 7,734,333,494,367,111

70 (17.5%) down, 330 (82.5%) to go.

[sje]

Draft 11 record number 71:
[d]rnbqkbnr/pppp1ppp/4p3/8/8/5N2/PPPPPPPP/RNBQKB1R w KQkq - 0 2[/d]
The perft(11) for the above is: 9,437,098,735,183,552

71 (17.75%) down, 329 (82.25%) to go.

[sje]

Draft 11 record number 72:
[d]rnbqkbnr/pppp1ppp/4p3/8/P7/8/1PPPPPPP/RNBQKBNR w KQkq - 0 2[/d]
The perft(11) for the above is: 9,038,950,832,777,449

72 (18%) down, 328 (82%) to go.

Mean time per draft 11 record: ca. 43 hours

[sje]

After 19 weeks, the run has produced 321,425 draft 8 records, about 39% of the total needed.

Mean time per draft 8 record: ca. 36 seconds

Due to a power outage five days ago, all data of draft seven or less was lost. If only the budget allowed for a UPS.

Most recent output:

Code: Select all

rnbqk1nr/ppppbppp/8/4p3/8/PP6/2PPPPPP/RNBQKBNR w KQkq - 1 3 9 5150421612315
rnbqk1nr/pppp1ppp/3b4/4p3/P7/3P3N/1PP1PPPP/RNBQKB1R b KQkq - 0 3 8 504518011072
rnb1kbnr/pppp1ppp/5q2/4p3/8/BP3P2/P1PPP1PP/RN1QKBNR b KQkq - 0 3 8 538758998566
rnb1kbnr/pppp1ppp/8/4p1q1/8/2NP3P/PPP1PPP1/R1BQKBNR b KQkq - 0 3 8 1089118079586
rnbqkb1r/pppp1ppp/5n2/4p3/8/PPP5/3PPPPP/RNBQKBNR b KQkq - 0 3 8 197822931197
rnbqk1nr/pppp1ppp/3b4/4p3/P2P4/7N/1PP1PPPP/RNBQKB1R b KQkq - 0 3 8 802521646000
rnbqkb1r/pppp1ppp/5n2/4p3/2P5/PP6/3PPPPP/RNBQKBNR b KQkq - 0 3 8 227512586756
rnbqkbnr/pppp2pp/5p2/4p3/P7/N3P3/1PPP1PPP/R1BQKBNR b KQkq - 0 3 8 520962677982
rnbqk1nr/pppp1ppp/3b4/4p3/P7/4P2N/1PPP1PPP/RNBQKB1R b KQkq - 0 3 8 827086849404
rnb1kbnr/pppp1ppp/5q2/4p3/5P2/BP6/P1PPP1PP/RN1QKBNR b KQkq - 0 3 8 729024463825

[Ajedrecista]

Hello:

After several days of calculation, I finally managed to calculate Perft(11) of 1.- f3, f6. I used JetChess 1.0.0.0 running in an Intel i5-760 (~ 2.8 GHz) with 4 GB of RAM, using 1 GB of hash for each Perft(10) count. It was impossible for me running the Perft(11) at once, so instead I ran 19 Perft(10) and added the results for getting Perft(11). I hope that everything is OK (I mean: no typos in numbers, correct FEN strings...).

Conditions were different for each Perft(10) calculation: I ran Perft(10) of 1.- f3, f6; 2.- * (* = Na3, Nc3, Nh3, a3 and a4) without running another Perft(10) at the same time; then I realized that I could run simultaneously two Perft(10) counts without hurting the performance (with the rest of positions except for 1.- f3, f6; 2.- d4 which I had to run all 19 Perft(9) values (which I added to get Perft(10) and then I deleted all the Perft(8) values I got)) of 1.- f3, f6; d4 because the computer freezed once or twice, and another time I ran out of time (I can not access to the i5 whenever I want).

Perft(11) of 1.- f3, f6 should look like this:

[d]rnbqkbnr/ppppp1pp/5p2/8/8/5P2/PPPPP1PP/RNBQKBNR w KQkq - 0 2

Code: Select all

  1  Nb1-a3   41794156721456
  2  Nb1-c3   55573541056355
  3  Ng1-h3   54797454951987
  4   a2-a3   35534716982746
  5   a2-a4   50028388474083
  6   b2-b3   47346825745074
  7   b2-b4   48072714223110
  8   c2-c3   55581388282559
  9   c2-c4   63072589712461
 10   d2-d3   96084529495520
 11   d2-d4  130944200886503
 12   e2-e3   92559716870913
 13   e2-e4  103757328933286
 14   g2-g3   47469016103682
 15   g2-g4   46963100928572
 16   h2-h3   34250790278448
 17   h2-h4   50498019974084
 18   f3-f4   52283241875269
 19  Ke1-f2   58690088930098

Total: 	  1165301810426206

1,165,301,810,426,206 (move pathes after 11 half moves).

Time: 185818.964 s (51:36:58.964).

This time of calculation (more than 51 hours) is not real time but CPU time as I understand it: for instance, if I run two Perft(10) simultaneously and both times of calculation are two hours, then real time are two hours, but CPU time are four hours (two hours plus two hours). I think that real time should be the half (more less) of these 51 hours. Fortunately, I could take advantage of simultaneous runs of Perft values. 1.- f3, f6; d4 was much more slower than the others due to I ran 19 Perft(9) instead of an unique Perft(10) which should increase the speed (although really the real time of calcultaion of this Perft(10) was a little less than five hours, in spite of the time I post, that is more than nine hours of CPU time). I am sure that if I can run the whole Perft(11) of 1.- f3, f6 at once, the time of calculation will be between 20 and 30 hours (a vague estimation) with the same conditions: Intel i5-760 and 1 GB of hash.

All the details (notepads with Perft(11), 19 Perft(10) and 361 Perft(9) (due to the output of JetChess) plus some chess diagrams, a PDF file and little more) can be downloaded from Mediafire:

Perft(11)_f3_f6.rar (974.79 KB)

I hope than sooner or later someone can check my results and confirm them. It was a great challenge for me (more than 1.1e+15 nodes counted) and I will not repeat something similar. Please enjoy it.

Regards from Spain.

Ajedrecista.

[sje]

Eventually the run will get to 1 f3 f6 draft 11. I'll guess that eight or more draft 12 results will come first, in about another month.

The next draft 11 record will likely be for 1 <X> e5 with <X>: {Na3 Nc3 Nf3 Nh3 a3 a4 b3 b4}.

[sje]

After 20 weeks, the run has produced 331,403 draft 8 records, about 40% of the total needed.

Mean time per draft 8 record: ca. 36 seconds

Most recent output:

Code: Select all

rnbqk1nr/ppppbppp/8/4p3/8/3P3N/PPPBPPPP/RN1QKB1R b KQkq - 2 3 8 439081704189
rnbqkbnr/pppp1p1p/6p1/4p3/P5P1/3P4/1PP1PP1P/RNBQKBNR b KQkq - 0 3 8 651089951521
rnbqkbnr/pppp1pp1/7p/4p3/5P2/2N1P3/PPPP2PP/R1BQKBNR b KQkq - 0 3 8 1022847249810
r1bqkbnr/pppp1ppp/n7/4p3/8/N2BP3/PPPP1PPP/R1BQK1NR b KQkq - 2 3 8 978461221829
rnbqkbnr/pppp1p1p/8/4p1p1/1P6/3P4/P1PQPPPP/RNB1KBNR b KQkq - 1 3 8 452030463425
rnbqkbnr/pppp1p1p/6p1/4p3/P7/3P3P/1PP1PPP1/RNBQKBNR b KQkq - 0 3 8 546602581638
r1bqkbnr/pppp1ppp/2n5/4p3/8/3P1N2/PPPQPPPP/RNB1KB1R b KQkq - 2 3 8 1169896133141
rnbqk1nr/ppppbppp/8/4p3/8/3PB2N/PPP1PPPP/RN1QKB1R b KQkq - 2 3 8 500525923972
rnbqkbnr/pppp2pp/5p2/4p3/3P4/P7/1PPQPPPP/RNB1KBNR b KQkq - 1 3 8 572130843907
rnbqkbnr/pppp1ppp/8/8/3p2P1/1P6/P1P1PP1P/RNBQKBNR b KQkq - 0 3 8 793609133355