Browse thread
OCaml troll on Slashdot
-
Karl Zilles
-
Oliver Bandel
-
Michael Vanier
-
Jon Harrop
-
Yoann Padioleau
- Jon Harrop
- Paul Argentoff
- Paul Argentoff
-
Yoann Padioleau
-
Jon Harrop
- Yoann Padioleau
-
Michael Vanier
- Richard Jones
-
Oliver Bandel
[
Home
]
[ Index:
by date
|
by threads
]
[ Message by date: previous | next ] [ Message in thread: previous | next ] [ Thread: previous | next ]
[ Message by date: previous | next ] [ Message in thread: previous | next ] [ Thread: previous | next ]
| Date: | -- (:) |
| From: | YANG Shouxun <yang.shx@f...> |
| Subject: | Re: [Caml-list] OCaml troll on Slashdot |
On Tuesday 15 March 2005 17:25, Richard Jones wrote: > On Mon, Mar 14, 2005 at 05:29:42PM -0800, Karl Zilles wrote: > > http://developers.slashdot.org/article.pl?sid=05/03/14/2258219 > > The problem was with his stupid implementation of a hash table: > > http://developers.slashdot.org/comments.pl?sid=142494&cid=11939530 > > (That comment is +5 informative) No. My experiments show that the OCaml implementation performs far better than the C++ implementation when the column and row get larger: C++ is compiled with -O3, not sure what is the better optimization level, while OCaml version (actually I used Sys.argv and references to the two parameters) is compiled with ocamlopt 4x4 c++ real 0m0.003s user 0m0.002s sys 0m0.002s 4x4 ocaml real 0m0.177s user 0m0.137s sys 0m0.001s 8x8 c++ real 0m8.703s user 0m7.000s sys 0m0.018s 8x8 ocaml real 0m0.747s user 0m0.485s sys 0m0.002s 12x12 c++ the process was killed by the OS 12x12 ocaml real 0m1.210s user 0m0.886s sys 0m0.001s