Browse Groups

• ## Statistical Results of «-l qs -fif 10 --flares-choice fcpro»

(1)
• NextPrevious
• Hi all, first of all - a note to Danny - your posts are much appreciated and I ll try to go over them more thoroughly and understand them Soon™, and see how
Message 1 of 1 , Dec 1 2:53 AM
View Source
Hi all,

first of all - a note to Danny - your posts are much appreciated and I'll try
to go over them more thoroughly and understand them Soon™, and see how easily
I can integrate it as an enhancement to Freecell Solver.

Back to the topic: here are the results of running
the freecell-solver-fc-pro-range-solver (segmented across 3,200 ten deal
segments and ran in parallel in 4 processes on my desktop machine, and 2 more
on my laptop) on the Microsoft 32,000 deals:

[QUOTE]
1 23
1 24
6 25
2 26
8 27
12 28
9 29
16 30
22 31
47 32
52 33
70 34
115 35
151 36
181 37
233 38
316 39
384 40
476 41
533 42
748 43
826 44
968 45
1166 46
1362 47
1373 48
1518 49
1688 50
1733 51
1775 52
1852 53
1733 54
1734 55
1611 56
1496 57
1388 58
1215 59
1031 60
882 61
665 62
604 63
492 64
392 65
290 66
218 67
179 68
117 69
83 70
66 71
58 72
29 73
24 74
18 75
13 76
8 77
5 78
1 79
2 83
1 84

[/QUOTE]

The statistics summary is:

[Q]
n is 31999, min is 23, max is 84
mode is 53 @ 1852, median is 50, mean is 52.5152, stdev is 7.23765
[/Q]

As one can see from the mfi-with-2-more-deals (which is essentially "-l qs"),
most of the statistics are lower, with a notable exception for the median
which remains 50 (*grunt*, *grunt*). That was for the record:

<<<<
n is 31999, min is 23, max is 92
mode is 53 @ 1729, median is 50, mean is 53.6379, stdev is 7.80832
>>>

But like I said the "-fif 10" made it consume a lot of time for a range of
deals.

The distribution of the deltas from the previous attempt is:

<<<<
1216 1
1024 2
851 3
739 4
614 5
525 6
448 7
356 8
250 9
227 10
172 11
139 12
94 13
77 14
62 15
59 16
33 17
23 18
26 19
12 20
10 21
10 22
6 23
7 24
3 25
1 26
1 27
1 29
2 30
1 32
>>>>

Here is the invocation I used:

<<<<<
#!/bin/bash

# This script parallelises the task of getting the results of a time-consuming
# flare-based scan using GNU parallel.

# To set up:
# mkdir -p ~/Arcs/FC_SOLVE_SUMMARIZE_RESULTS/ on sh.
# mkdir -p ~/Arcs/FC_SOLVE_SUMMARIZE_RESULTS/ on lap.
# sshfs sh:Arcs/FC_SOLVE_SUMMARIZE_RESULTS/ ~/Arcs/FC_SOLVE_SUMMARIZE_RESULTS/
on lap.
# ../Tatzer -l ci7b --prefix="\$HOME/apps/fcs" ; make ; make install on sh.
# ../Tatzer -l c2b --prefix="\$HOME/apps/fcs" ; make ; make install on lap.

"\$HOME"/apps/fcs/bin/freecell-solver-fc-pro-range-solve \
"\\$(({}*10+1))" "\\$((({}+1)*10))" 1 \
-l qs -fif 10 --flares-choice fcpro \> \
"\$HOME"/Arcs/FC_SOLVE_SUMMARIZE_RESULTS/mfi-with-two-more-scans-flares-choice-fcpro.fc-pro-dump__\\$\(printf
\"%06d\" \"{}\"\).txt \; echo "Finished {}"
>>>>>

Regards,

Shlomi Fish

--
-----------------------------------------------------------------
Shlomi Fish http://www.shlomifish.org/
"The Human Hacking Field Guide" - http://shlom.in/hhfg

Real programmers use a nice editor and a nice programming language and get it
done in less than O(N!).
— vanguard on Freenode’s ##programming