On Wednesday, 24 September 2014 at 19:38:37 UTC, bearophile wrote:
Joakim:
I wonder why they found Haskell to be so slow, I thought it
was compiled.
The first reason for the performance of programs is how much
care the programmer has to write a fast program, secondly how
good the chosen algorithms are, and only then at a third place
there are language implementations.
So Haskell programmers don't care about fast programs and don't
know how to choose good algorithms? ;)
So the short answer to your question is that most of the
Haskell programs of Rosettacode are not written for
performance. The corollary is that performance comparison is
silly/bogus.
I understand that Rosettacode is not geared towards performance,
but that's why it makes a good sample for the authors of
idiomatic code in each language, albeit potentially skewed by the
skill of the volunteers contributing. What's interesting is that
the performance results for most of the other languages are about
what I'd expect, except for Haskell.
Often the faster programs of Rosettacode are in... guess a
language? Yes, in D. Because I have written also performance
conscious D entries (sometimes I have added various D
solutions, at various levels of performance, succinctness,
safety & correctness).
Yes, I've seen your samples and different versions on that site
before. Good to see you put so much time into publicizing D, but
I wonder how well-known that site is, as I never heard of it till
you mentioned your involvement with it here.
