Advantages of a large number of threads cf other approaches?

[Joe Armstrong]

On Tue, 17 Feb 2004, Joachim Durchholz wrote:

> Joe Armstrong wrote:
> 
> > Interestingly  Erlang  and  Apache  perform equally  when  both  are 
> > unloaded  - this is  hardly surprising  since the  heavy stuff  in
> > the Erlang I/O  routines is  all written  in C and  the programs  are
> > "BIF bound" -  but under conditions  of massive overload  the story
> > is very different.
> > 
> > To see how things shape up under massive overload see:
> > 
> >   http://www.sics.se/~joe/apachevsyaws.html
> > 
> > In this  experiment Apache crashed when  subject to a  load of about 
> > 4000  parallel sessions  - the  Erlang web  server (yaws)  was
> > happily ticking along at 80,000 parallel sessions.
> 
> That's interesting.
> 
> Joe, would you or Ali be interested in running the same test with a
> Linux 2.6 kernel? They implemented an O(1) scheduler there, and it would
> help to determine whether Apache's crash is due to OS process issues or
> due to Apache.

  Right now I don't have any time to re-do the tests.

  My guess is that the problem  lies in the kernel. At 4000+ processes
both Java and  C++ started behaving in a wierd manner  - I don't think
this was a co-incidence.

/Joe

> 
> Regards,
> Jo
> --
> Currently looking for a new job.
>