Hi Andreas, <div>For some reason, lists:usort is your friend again! </div><div>Looking into the code deeper, one line can give u huge boost!</div><div>You compare the unit values with digits with </div><div><div> (length(UnitValues) == 9)</div>
<div> and (gb_sets:from_list(UnitValues) == gb_sets:from_list(digits())).</div><div><br></div><div>I understand (correct me if i'm wrong) u check first the length to make sure it's 9 digits then u convert to sets to make sure they r unique and compare them. But if you think about it, if the unique list of values is of length 9, u don't need more comparisons!</div>
<div><br></div><div>so u can actually just do </div><div>(length(lists:usort(UnitValues)==9)</div><div><br></div><div>before the change: </div><div><div>sudoku:print_results("top95.txt").</div><div>Solved 95 of 95 puzzles from top95.txt in 1.937175 secs (49.04 Hz)</div>
<div> (210794 total eliminations, avg 2218.88, median 1536, max 10664, min 648).</div></div><div><br></div><div>after the change : </div><div><div>sudoku:print_results("top95.txt").</div><div>Solved 95 of 95 puzzles from top95.txt in 0.382337 secs (248.47 Hz)</div>
<div> (45002 total eliminations, avg 473.71, median 479, max 518, min 427).</div></div><div><br></div><br><div class="gmail_quote">On Mon, Mar 28, 2011 at 3:32 PM, Andreas Pauley <span dir="ltr"><<a href="mailto:apauley@gmail.com">apauley@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">Hi Tristan,<br>
<br>
I tried the code in your repo, but it wasn't significantly faster on my machine.<br>
<br>
It was interesting though to go through your refactorings.<br>
I see you changed the signature for solve so that the empty puzzle and<br>
the list of square names are passed in from the start.<br>
Maybe I should try this approach in conjunction with the conversion to atoms...<br>
<br>
Regards,<br>
<font color="#888888">Andreas<br>
</font><div><div></div><div class="h5"><br>
On Sat, Mar 26, 2011 at 10:56 PM, Tristan Sloughter<br>
<<a href="mailto:tristan.sloughter@gmail.com">tristan.sloughter@gmail.com</a>> wrote:<br>
> With a few refactorings (<a href="https://github.com/tsloughter/sudoku-in-erlang" target="_blank">https://github.com/tsloughter/sudoku-in-erlang</a> --<br>
> note I used ewl_plists just to make the code simpler:<br>
> <a href="https://github.com/erlware/erlware/tree/master/lib/ewlib" target="_blank">https://github.com/erlware/erlware/tree/master/lib/ewlib</a>) and compile with<br>
> native, hipe o3 it beats python's:<br>
> $ ./sudoku.py<br>
> All tests pass.<br>
> Solved 50 of 50 puzzles from easy50.txt in 0.529119 secs (94.50 Hz)<br>
> (33059 total eliminations, avg 661.00, median 648, max 830, min 648).<br>
> Solved 95 of 95 puzzles from top95.txt in 3.889944 secs (24.42 Hz)<br>
> (221997 total eliminations, avg 2336.00, median 1492, max 11512, min 648).<br>
> Solved 11 of 11 puzzles from hardest.txt in 0.157520 secs (69.83 Hz)<br>
> (9436 total eliminations, avg 857.00, median 817, max 1198, min 648).<br>
> $ ./sudoku<br>
> All tests passed :-)<br>
> Solved 50 of 50 puzzles from easy50.txt in 0.556385 secs (89.87 Hz)<br>
> (92538 total eliminations, avg 1850.76, median 1811, max 2628, min 1767).<br>
> Solved 95 of 95 puzzles from top95.txt in 3.593476 secs (26.44 Hz)<br>
> (901201 total eliminations, avg 9486.33, median 6267, max 56820, min<br>
> 1792).<br>
> Solved 11 of 11 puzzles from hardest.txt in 0.162107 secs (67.86 Hz)<br>
> (33653 total eliminations, avg 3059.36, median 3023, max 5346, min 1786).<br>
> I ran a number of times, those are pretty much what it always comes out to.<br>
> But I haven't really tried a full refactoring, I think a number of things<br>
> can be reduced to less iterations.<br>
><br>
> Tristan<br>
> On Sat, Mar 26, 2011 at 3:39 PM, Ahmed Omar <<a href="mailto:spawn.think@gmail.com">spawn.think@gmail.com</a>> wrote:<br>
>><br>
>> BTW, beside looking into reducing eleminations (which requires deep dive),<br>
>> if u do some time profiling using eprof u will find that most of the time<br>
>> spent actually in the cross functions<br>
>> sudoku:'-cross/2-lc$^1/1-1-'/4 96558660 37.46<br>
>> 35024918 [ 0.36]<br>
>><br>
>> On Sat, Mar 26, 2011 at 9:33 PM, Ahmed Omar <<a href="mailto:spawn.think@gmail.com">spawn.think@gmail.com</a>> wrote:<br>
>>><br>
>>> actually instead of doing<br>
>>><br>
>>> NonUniquePeers = shallow_flatten([S || S <- units(Square)]),<br>
>>><br>
>>><br>
>>> PeerSet = sets:from_list(NonUniquePeers),<br>
>>><br>
>>><br>
>>> PeersWithSelf = sets:to_list(PeerSet),<br>
>>><br>
>>><br>
>>><br>
>>><br>
>>><br>
>>><br>
>>><br>
>>> can't u just do<br>
>>> PeersWithSelf = lists:usort(NonUniquePeers).<br>
>>><br>
>>><br>
>>> ?<br>
>>><br>
>>><br>
>>><br>
>>><br>
>>><br>
>>><br>
>>><br>
>>> On Sat, Mar 26, 2011 at 9:14 PM, Evans, Matthew <<a href="mailto:mevans@verivue.com">mevans@verivue.com</a>><br>
>>> wrote:<br>
>>>><br>
>>>> Good call....gb_sets should be faster. Compiled to native it runs faster<br>
>>>> still<br>
>>>><br>
>>>> 9> sudoku:print_results("top95.txt", "\n").<br>
>>>> Solved 95 of 95 puzzles from top95.txt in 0.810156 secs (117.26 Hz)<br>
>>>> (901201 total eliminations, avg 9486.33, median 6267, max 56820, min<br>
>>>> 1792).<br>
>>>> ok<br>
>>>><br>
>>>> ________________________________________<br>
>>>> From: Ahmed Omar [<a href="mailto:spawn.think@gmail.com">spawn.think@gmail.com</a>]<br>
>>>> Sent: Saturday, March 26, 2011 3:08 PM<br>
>>>> To: Andreas Pauley<br>
>>>> Cc: Evans, Matthew; <a href="mailto:erlang-questions@erlang.org">erlang-questions@erlang.org</a><br>
>>>> Subject: Re: [erlang-questions 20] Re: A sudoku solver in Erlang<br>
>>>> compared to Python<br>
>>>><br>
>>>> using gb_sets instead of sets could decrease eliminations a bit and give<br>
>>>> u some boost. However, i didn't dive deeper into the code or the algorithm.<br>
>>>><br>
>>>> On Sat, Mar 26, 2011 at 9:41 AM, Andreas Pauley<br>
>>>> <<a href="mailto:apauley@gmail.com">apauley@gmail.com</a><mailto:<a href="mailto:apauley@gmail.com">apauley@gmail.com</a>>> wrote:<br>
>>>> Thanks, I've changed the compile options and this definitely makes it<br>
>>>> somewhat faster:<br>
>>>><br>
>>>> <a href="https://github.com/apauley/sudoku-in-erlang/commit/b7ce2abb6ca013850ed8f3e8fd7f5f6be7004cbb" target="_blank">https://github.com/apauley/sudoku-in-erlang/commit/b7ce2abb6ca013850ed8f3e8fd7f5f6be7004cbb</a><br>
>>>><br>
>>>> Strangely the native flag is not documented here:<br>
>>>> <a href="http://www.erlang.org/doc/man/compile.html#file-2" target="_blank">http://www.erlang.org/doc/man/compile.html#file-2</a><br>
>>>><br>
>>>> The more interesting improvement would be to decrease the number of<br>
>>>> eliminations performed, but for that I'll have to go deep into the<br>
>>>> code :-)<br>
>>>><br>
>>>><br>
>>>> On Sat, Mar 26, 2011 at 12:16 AM, Evans, Matthew<br>
>>>> <<a href="mailto:mevans@verivue.com">mevans@verivue.com</a><mailto:<a href="mailto:mevans@verivue.com">mevans@verivue.com</a>>> wrote:<br>
>>>> > Without going deep into the code, one thing to try is compile with the<br>
>>>> > native flag (running the tests from the VM shell here):<br>
>>>> ><br>
>>>> > Without native set:<br>
>>>> > 18> c(sudoku).<br>
>>>> > {ok,sudoku}<br>
>>>> > 19> sudoku:print_results("top95.txt", "\n").<br>
>>>> > Solved 95 of 95 puzzles from top95.txt in 2.038825 secs (46.60 Hz)<br>
>>>> > (922678 total eliminations, avg 9712.40, median 6596, max 55370, min<br>
>>>> > 1797).<br>
>>>> > ok<br>
>>>> ><br>
>>>> ><br>
>>>> > With native set:<br>
>>>> > 20> c(sudoku,[native]).<br>
>>>> > {ok,sudoku}<br>
>>>> > 21> sudoku:print_results("top95.txt", "\n").<br>
>>>> > Solved 95 of 95 puzzles from top95.txt in 1.613416 secs (58.88 Hz)<br>
>>>> > (922678 total eliminations, avg 9712.40, median 6596, max 55370, min<br>
>>>> > 1797).<br>
>>>> ><br>
>>>> ><br>
>>>> ><br>
>>>> > -----Original Message-----<br>
>>>> > From:<br>
>>>> > <a href="mailto:erlang-questions-bounces@erlang.org">erlang-questions-bounces@erlang.org</a><mailto:<a href="mailto:erlang-questions-bounces@erlang.org">erlang-questions-bounces@erlang.org</a>><br>
>>>> > [mailto:<a href="mailto:erlang-questions-bounces@erlang.org">erlang-questions-bounces@erlang.org</a><mailto:<a href="mailto:erlang-questions-bounces@erlang.org">erlang-questions-bounces@erlang.org</a>>]<br>
>>>> > On Behalf Of Andreas Pauley<br>
>>>> > Sent: Friday, March 25, 2011 8:15 AM<br>
>>>> > To: <a href="mailto:erlang-questions@erlang.org">erlang-questions@erlang.org</a><mailto:<a href="mailto:erlang-questions@erlang.org">erlang-questions@erlang.org</a>><br>
>>>> > Subject: [erlang-questions 4] A sudoku solver in Erlang compared to<br>
>>>> > Python<br>
>>>> ><br>
>>>> > Hi all,<br>
>>>> ><br>
>>>> > In my quest to learn Erlang I've translated Peter Norvig's sudoku<br>
>>>> > solver into Erlang:<br>
>>>> > <a href="https://github.com/apauley/sudoku-in-erlang" target="_blank">https://github.com/apauley/sudoku-in-erlang</a><br>
>>>> ><br>
>>>> > For comparison, my slightly modified version of Norvig's Python code<br>
>>>> > can be found here:<br>
>>>> > <a href="https://github.com/apauley/sudoku-by-norvig" target="_blank">https://github.com/apauley/sudoku-by-norvig</a><br>
>>>> ><br>
>>>> > I like the pattern matching, it was fun to do the entire solution<br>
>>>> > without a single if or case statement :-)<br>
>>>> > Something that bothers me though is that the Python solution seems to<br>
>>>> > be faster, even after I've made the Erlang solution use multiple<br>
>>>> > processors.<br>
>>>> ><br>
>>>> > In order to compare the two solutions I counted the number of<br>
>>>> > eliminations each performed.<br>
>>>> > The eliminate function is the core of the solution, for example<br>
>>>> > assigning a single value to a square is implemented as the elimination<br>
>>>> > of all other values.<br>
>>>> ><br>
>>>> > With the Erlang implementation I get:<br>
>>>> > sudoku-in-erlang$ ./sudoku<br>
>>>> > All tests passed :-)<br>
>>>> > Solved 50 of 50 puzzles from easy50.txt in 2.890978 secs (17.30 Hz)<br>
>>>> > (93403 total eliminations, avg 1868.06, median 1810, max 2517, min<br>
>>>> > 1770).<br>
>>>> > Solved 95 of 95 puzzles from top95.txt in 22.004369 secs (4.32 Hz)<br>
>>>> > (922678 total eliminations, avg 9712.40, median 6596, max 55370, min<br>
>>>> > 1797).<br>
>>>> > Solved 11 of 11 puzzles from hardest.txt in 0.851678 secs (12.92 Hz)<br>
>>>> > (32339 total eliminations, avg 2939.91, median 2894, max 4779, min<br>
>>>> > 1781).<br>
>>>> ><br>
>>>> > And with the Python implementation:<br>
>>>> > sudoku-by-norvig$ ./sudoku.py<br>
>>>> > All tests pass.<br>
>>>> > Solved 50 of 50 puzzles from easy50.txt in 0.792008 secs (63.13 Hz)<br>
>>>> > (33059 total eliminations, avg 661.00, median 648, max 830, min 648).<br>
>>>> > Solved 95 of 95 puzzles from top95.txt in 5.903875 secs (16.09 Hz)<br>
>>>> > (221997 total eliminations, avg 2336.00, median 1492, max 11512, min<br>
>>>> > 648).<br>
>>>> > Solved 11 of 11 puzzles from hardest.txt in 0.237532 secs (46.31 Hz)<br>
>>>> > (9436 total eliminations, avg 857.00, median 817, max 1198, min 648).<br>
>>>> ><br>
>>>> > So according to the stats above, the Python solution performs less<br>
>>>> > computations when given exactly the same input.<br>
>>>> > The Erlang code is as close to the Python as I could make it, I've<br>
>>>> > done more or less a direct translation of the algorithms used.<br>
>>>> ><br>
>>>> > I suspect that there are some lazy evaluation happening in the Python<br>
>>>> > version, possibly generators, although I haven't pinpointed it yet.<br>
>>>> ><br>
>>>> > How can I improve my Erlang code in this solution?<br>
>>>> ><br>
>>>> > Kind regards,<br>
>>>> > Andreas<br>
>>>> > _______________________________________________<br>
>>>> > erlang-questions mailing list<br>
>>>> > <a href="mailto:erlang-questions@erlang.org">erlang-questions@erlang.org</a><mailto:<a href="mailto:erlang-questions@erlang.org">erlang-questions@erlang.org</a>><br>
>>>> > <a href="http://erlang.org/mailman/listinfo/erlang-questions" target="_blank">http://erlang.org/mailman/listinfo/erlang-questions</a><br>
>>>> ><br>
>>>> _______________________________________________<br>
>>>> erlang-questions mailing list<br>
>>>> <a href="mailto:erlang-questions@erlang.org">erlang-questions@erlang.org</a><mailto:<a href="mailto:erlang-questions@erlang.org">erlang-questions@erlang.org</a>><br>
>>>> <a href="http://erlang.org/mailman/listinfo/erlang-questions" target="_blank">http://erlang.org/mailman/listinfo/erlang-questions</a><br>
>>>><br>
>>>><br>
>>>><br>
>>>> --<br>
>>>> Best Regards,<br>
>>>> - Ahmed Omar<br>
>>>> <a href="http://nl.linkedin.com/in/adiaa" target="_blank">http://nl.linkedin.com/in/adiaa</a><br>
>>>> Follow me on twitter<br>
>>>> @spawn_think<<a href="http://twitter.com/#!/spawn_think" target="_blank">http://twitter.com/#!/spawn_think</a>><br>
>>>><br>
>>><br>
>>><br>
>>><br>
>>> --<br>
>>> Best Regards,<br>
>>> - Ahmed Omar<br>
>>> <a href="http://nl.linkedin.com/in/adiaa" target="_blank">http://nl.linkedin.com/in/adiaa</a><br>
>>> Follow me on twitter<br>
>>> @spawn_think<br>
>><br>
>><br>
>><br>
>> --<br>
>> Best Regards,<br>
>> - Ahmed Omar<br>
>> <a href="http://nl.linkedin.com/in/adiaa" target="_blank">http://nl.linkedin.com/in/adiaa</a><br>
>> Follow me on twitter<br>
>> @spawn_think<br>
>><br>
>> _______________________________________________<br>
>> erlang-questions mailing list<br>
>> <a href="mailto:erlang-questions@erlang.org">erlang-questions@erlang.org</a><br>
>> <a href="http://erlang.org/mailman/listinfo/erlang-questions" target="_blank">http://erlang.org/mailman/listinfo/erlang-questions</a><br>
>><br>
><br>
><br>
</div></div></blockquote></div><br><br clear="all"><br>-- <br>Best Regards,<br>- Ahmed Omar<div><a href="http://nl.linkedin.com/in/adiaa" target="_blank">http://nl.linkedin.com/in/adiaa</a></div><div>Follow me on twitter</div>
<div><a href="http://twitter.com/#!/spawn_think" target="_blank">@spawn_think</a></div><br>
</div>