Perft(13), final fifth

[sje]

After 46 weeks:

The run has produced 702,648 draft 8 records, about 85% of the total needed. The average draft 8 record generation time is about 39.6 seconds.

There are now 257 draft 11 records, 64.25% of the 400 needed. The average draft 11 record generation time is about 30.1 hours.
http://dl.dropbox.com/u/31633927/Perft/Perft13/draft11

There are now 10 draft 12 records, 50% of the 20 needed. The average draft 12 record generation time is about 32.2 days.
http://dl.dropbox.com/u/31633927/Perft/Perft13/draft12

Uptime:

Code: Select all

10 days,  8:56

Recent output:

Code: Select all

rnb1kbnr/pppp1ppp/8/4p3/Q1P4q/6P1/PP1PPP1P/RNB1KBNR b KQkq - 0 3 8 1671605666120
rnbqk1nr/ppppbppp/8/1B2p3/8/2P1P3/PP1P1PPP/RNBQK1NR b KQkq - 2 3 8 669297224640
rnbqk1nr/pppp1ppp/4p3/8/1b2P3/5QP1/PPPP1P1P/RNB1KBNR b KQkq - 2 3 8 1670598207567
rnbqkbnr/p1p1pppp/1p1p4/8/6PP/7R/PPPPPP2/RNBQKBN1 b Qkq - 1 3 8 368894453423
rnb1kbnr/pppp1ppp/4p3/8/4P1Pq/3P4/PPP2P1P/RNBQKBNR b KQkq - 0 3 8 1364700734462
rnb1kbnr/pppqpppp/7B/3p4/3P4/4P3/PPP2PPP/RN1QKBNR b KQkq - 0 3 8 1240900162247
rnbqkbnr/ppp2ppp/8/3pp2Q/8/4P2P/PPPP1PP1/RNB1KBNR b KQkq - 0 3 8 2484204132415
rnbqk1nr/ppppbppp/B7/4p3/8/2P1P3/PP1P1PPP/RNBQK1NR b KQkq - 2 3 8 709686820630
rnbqkbnr/p1p1pppp/1p1p4/6P1/7P/8/PPPPPP2/RNBQKBNR b KQkq - 0 3 8 229197593179
rnbqk1nr/ppppbppp/8/4p3/8/2P1P3/PP1PNPPP/RNBQKB1R b KQkq - 2 3 8 372359988871

[sje]

Due to heat problems, the run has taken a short vacation. It has since resumed and within a day or so will enter the last seventh of the draft eight record generation.

I'm making a guess that the draft nine count for this ply four position
[d]
will be the highest of all the draft nine counts in the run. I have calculated the value in a separate run on a separate machine using a 56 bit hash and the result is: 135,572,854,768,241. (Three hours, 45 minutes.) Perhaps someone might try to confirm or deny this result with their own calculation.

[ibid]

I have calculated the value in a separate run on a separate machine using a 56 bit hash and the result is: 135,572,854,768,241. (Three hours, 45 minutes.) Perhaps someone might try to confirm or deny this result with their own calculation.

Confirmed.

-paul

[sje]

I have calculated the value in a separate run on a separate machine using a 56 bit hash and the result is: 135,572,854,768,241. (Three hours, 45 minutes.) Perhaps someone might try to confirm or deny this result with their own calculation.

Confirmed.

-paul

Thanks. I re-ran with a 120 bit hash and got the same result.

The new estimated completion date is August 18th, 54 days from today.

[sje]

The current maximum ply five, draft eight:
[d]
The perft(8) for the above is 6,502,542,258,889.


The current maximum ply four, draft nine count:
[d]
The perft(9) for the above is 174,561,830,680,905.


The current maximum ply three, draft ten count:
[d]
The perft(10) for the above is 1,914,771,616,164,660.


The current maximum ply two, draft eleven count:
[d]
The perft(11) for the above is 17,344,682,849,928,475.

[sje]

After 47 weeks:

The run has produced 708,411 draft 8 records, about 86% of the total needed. The average draft 8 record generation time is about 40.1 seconds.

There are now 263 draft 11 records, 65.75% of the 400 needed. The average draft 11 record generation time is about 30.0 hours.
http://dl.dropbox.com/u/31633927/Perft/Perft13/draft11

There are now 10 draft 12 records, 50% of the 20 needed. The average draft 12 record generation time is about 32.9 days.
http://dl.dropbox.com/u/31633927/Perft/Perft13/draft12

Uptime:

Code: Select all

5 days,  2:36

Recent output:

Code: Select all

rnbqkbnr/ppp1ppp1/8/3p3p/3PQ3/8/PPP1PPPP/RNB1KBNR b KQkq - 1 3 8 1392355398818
rnb1kbnr/pppp1ppp/8/4p3/7q/6P1/PPPPPPBP/RNBQK1NR b KQkq - 3 3 8 1286582163976
rnbqkbnr/ppp1ppp1/8/1Q1p3p/3P4/8/PPP1PPPP/RNB1KBNR b KQkq - 1 3 8 173100917885
rnb1kbnr/pppp1ppp/8/4p1q1/2P4P/8/PP1PPPP1/RNBQKBNR b KQkq - 0 3 8 948800336071
rnbqkb1r/ppp1pppp/7n/3p4/4P1Q1/8/PPPPKPPP/RNB2BNR b kq - 3 3 8 919960007349
rnbqkbnr/pppp1p1p/6p1/4p3/2PP1B2/8/PP2PPPP/RN1QKBNR b KQkq - 1 3 8 1260719345582
rnbqkbnr/ppp2ppp/4p3/3p4/8/3BP3/PPPP1PPP/RNBQK1NR b KQkq - 1 3 8 1546324543062
rnbqkbnr/ppp1p1pp/5p2/3p4/6P1/4P3/PPPP1PBP/RNBQK1NR b KQkq - 0 3 8 430297802188
rnb1kbnr/pppp1ppp/8/4p3/6Bq/6P1/PPPPPP1P/RNBQK1NR b KQkq - 3 3 8 767942809027
rnbqkbnr/ppp1ppp1/8/3p1Q1p/3P4/8/PPP1PPPP/RNB1KBNR b KQkq - 1 3 8 1256547343135

[sje]

Due to the unusual summer heat here in New England, the perft(13) run is being occasionally suspended during warmer (i.e., 90+ °F) days.

To help check the results of the perft(13) run, I have been making several parallel sub-calculations on secondary machines. One of these is a perft(11) run in progress for the position after 1 e4 e5; I guess that this will be the largest of all perft(11) subtotals for perft(13). This calculation has produced 7 of the 29 draft 10 subtotals so far; I hope to have the perft(11) number in a week or so.

[sje]

For the perft(11) of the position after 1 e4 e5, here are the first 7 of the 29 draft 10 results:

Code: Select all

Bd3 573,419,513,041,512
Bb5 647,063,600,292,176
Ba6 722,426,373,889,527
Bc4 980,479,287,343,390
Be2 517,614,317,370,382
Ke2 311,726,431,074,729
Na3 661,892,146,915,721

[sje]

For the perft(11) of the position after 1 e4 e5, here are the rest of the draft 10 knight moves:

Code: Select all

Nc3 862,976,539,962,908
Ne2 362,870,123,116,214
Nf3 634,862,113,468,243
Nh3 682,529,240,384,687

In progress: 4 queen moves, 14 pawn moves.

[sje]

For the perft(11) of the position after 1 e4 e5, here are the four draft 10 queen moves:

Code: Select all

Qe2 809,254,881,641,774
Qf3 1,461,016,305,993,700
Qg4 1,424,254,566,341,054
Qh5 1,160,663,490,726,049

In progress: 14 pawn moves