Comment by SwellJoe
15 years ago
I was impressed by the brevity of the C solution using OpenMP (comment #81), especially compared to significantly more verbose Erlang and Haskell examples. I kept reading it, thinking, "Where's the fork?" and seeing nothing but a simple for...and then remembered that # is a preprocessor directive, rather than a comment, in C, and then googled for "parallel omp for". So, I actually learned something today from that ridiculous thread: http://en.wikipedia.org/wiki/OpenMP
Also, the Perl 6 one-liner is sort of laughably cute . I'm not sure I believe it would actually work, but if it does, the conciseness of Perl 6 is even higher than I'd realized; dropping about 9 words from the Perl 5 one-liner (which is already very small). But, I learned about the hyper operators in Perl 6, which is also cool: http://perl6advent.wordpress.com/2009/12/06/day-6-going-into...
In short, that was awesome reading over my morning tea.
There's a few short implementations of this algorithm on CodeGolf.SE: http://codegolf.stackexchange.com/questions/2722/implement-s...
I'm relieved to note that Haskell and Erlang don't have to be ridiculously verbose. There's even an Erlang one-liner that is competitive with the Perl and Ruby implementations for brevity. I was amused to note that Perl is still the shortest implementation, and could be shorter (though the frighteningly short Perl 6 implementation mentioned in the 4chan thread is either wrong, or the implementation of the <<. operator is still unfinished in Rakudo Star, because it does not actually do the job; I don't understand the hyperoperators well enough yet to know which is true).
significantly more verbose Erlang and Haskell examples
There are less verbose (but also less well-formed) solutions in these languages. Here's a working, hacky but still readable, 3-line one I just banged out for Erlang: http://pastebin.com/eZckJp9p
... and a version that trades a line for calling an actual function named sleep(), given there are rules to this silly game http://pastebin.com/heQ43K0E ...
Also, the Perl 6 one-liner is sort of laughably cute . I'm not sure I believe it would actually work
I don't think any current implementation supports auto-parallelization of hyperops. Regardless, hyper is really only spec'd to remove the result-ordering requirements so that it may optimize it (I wouldn't be surprised if was executed in chunks in slightly more sophisticated implementations.)
From my reading of the docs, it (>>.) doesn't seem to require parallelization in the implementation, at all...it just allows it and requires code to be parallel safe, but doesn't impose any parallel requirements on the Perl 6 implementation. So, something like that Perl one-liner that would require parallelism to even work at all simply isn't valid. It might be possible to build a Perl 6 that would exhibit the right behavior, but it would be unspecified behavior.
So, I think that tiny little implementation is bunk. I'd be curious to see what the actual smallest Perl 6 implementation would look like, or if it would be the smallest Perl 5 version except using "say" to shave off two characters.
(>>.) doesn't seem to require parallelization in the implementation, at all...it just allows it and requires code to be parallel safe
That's pretty much what I said.
I'd be curious to see what the actual smallest Perl 6 implementation would look like
If you want to guarantee it will work, then it will be about as long as a Perl 5 version. Every brief way to parallelize leaves a lot of the semantics up to the implementation. There's a good reason for that: this is a concurrent problem, but the hyperops are meant for parallelizable problems.
3 replies →