Perft(12) re-run

[sje]

The perft(12) re-run continues on the 3.4 GHz i7-2600 machine.

After 222 hours wall time:

8 of 20 perft(11)
206 of 400 perft(10)
3,494 of 5,362 perft(9)
50,150 of 72,078 perft(8) (ca. 69.6% of total)

Current total time estimate: 319 hours (ca. 13.3 days)

The eight (Na3 Nc3 Nf3 Nh3 a3 a4 b3 b4) perft(11) subtotals generated at this point all match established results.

The above estimate is about 21.9 times larger than the time needed to do perft(11) on the same machine with the same program.

The 13.3 day estimate compares favorably to the 118 days it took to do perft(12) on a motley assortment of machines.

[sje]

The perft(12) re-run has completed and all the results were correct. The run took 13.5 days, slightly longer than expected. This is the first time that perft(12) was calculated on a single machine and with a single, uninterrupted run.

The run was a good test of the new hardware.

I'll be making some adjustments and doing a few short tests before starting perft(13) on the new machine.