On Thu, 21 Mar 2002, Shlomi Fish wrote:
> The following run of the range parallel solver:
>
> freecell-solver-range-parallel-solve 1 32000 1 ...
>
> Resulted in a run-time of 73 minutes[1] and totalled 26,765,593 iteration,
> which is slightly better than the corrseponding run without using
> Soft-Threads.
>
> [1] - On a Pentium III 667 MHz.
% time patsolve -fS -N1 32000 -q
1012.446u 23.412s 32:12.61 53.5%
That is, 17.2 minutes of runtime which took 32 min on my 666 MHz Alpha
running one other process (thus 53.5% CPU utilization).
Patsolve is single-threaded, and deterministic. It's just clever
about which branch to follow; there are about 11 weighted heuristics,
and the weighting was chosen by genetic algorithm training.
Bill Raymond's solver has hundreds of heuristics...
Dr. Tom Holroyd
"I am, as I said, inspired by the biological phenomena in which
chemical forces are used in repetitious fashion to produce all
kinds of weird effects (one of which is the author)."
-- Richard Feynman, _There's Plenty of Room at the Bottom_
Received on Thu Mar 21 2002 - 18:16:15 IST