[Shootout-list] Added C# -- But a problem...

[Isaac Gouy]

> All C# tests complete in 0.06 user time, which leads me to believe we
> aren't measuring the component of the system actually doing the work

For Ackermann, measurements on Gentoo Linux using the fork/exec/waitpid
technique seem to be directly comparable to Java:

      Java      Mono C#
n=4   0.13      0.11
n=5   0.13      0.11
n=6   0.13      0.12
n=7   0.15      0.13
n=8   0.21      0.19
n=9             0.49
n=10            4.98


Mono measurements:
           $cuser  $csys  ($cuser+$csys)
   Ack(3,4): 125
   Mono C#   0.1   0.01   0.11

   Ack(3,5): 253
   Mono C#   0.11   0   0.11

   Ack(3,6): 509
   Mono C#   0.1   0.02   0.12

   Ack(3,7): 1021
   Mono C#   0.11   0.02   0.13

   Ack(3,8): 2045
   Mono C#   0.19   0   0.19

   Ack(3,9): 4093
   Mono C#   0.48   0.01   0.49

   Ack(3,10): 8189
   Mono C#   4.96   0.02   4.98


Perl script:

   use BSD::Resource qw(times);
   my($cu0, $cs0) = (BSD::Resource::times)[2,3];

   unless (defined ($pid = fork)) {die "cannot fork: $!"; } 
   unless ($pid) { exec 'mono ackermann.exe 4'; exit; } 
   waitpid($pid, 0);

   my($cuser, $csys) = (BSD::Resource::times)[2,3];
   $cuser -= $cu0;
   $csys -= $cs0;
   $t = $cuser + $csys;
   print "Mono C#   $cuser   $csys   $t\n\n";



		
__________________________________
Do you Yahoo!?
Yahoo! Mail is new and improved - Check it out!
http://promotions.yahoo.com/new_mail