[oclug] Programming Wars 2.0 - September 2001

[dlinford at vcn.bc.ca]

Vic Gedris oclug at lists.oclug.on.ca 
> On Fri, 3 Aug 2001 dlinford at vcn.bc.ca wrote:
> 
> > But that would make it more interesting, especially regarding
> > which languages can deal with "the real world" well, in Linux!
> > That's also a _very_ soft real time problem, with the hardest
> > part being how to specify the requirement.
> >
> > Some of the issues this tease at make me tempted, but:
> >
> 
> >       c) I'm not into learning enough of "threads" to make
> >          the sort of use one should to exploit the
> >          techniques well.
> 
> It would have been interesting to have a multi-processor machine to test
> these on.  

I guess, but unlikely more so than with a modern uniprocessor. 

> .  Do linux threads take advantage of multiple CPUs, or is it
> just processes?  

Threads came about less for multiple CPUs, than for other reasons. 
Unfortunately, I look at them for using multiple CPUs from before 
they really came about. "Same" thing, different angle.  :-(

>   Would multiple threads even be useful in the
> calculation of primes?

Would depend on how one went about it... But there's more to the 
task than calculation, isn't there?  :-)

It's more the issues of "high performance", but and this 
problem happens to be an interesting excuse. It's the 
techniques that are cool, and aren't as annoyingly difficult 
to use in unix than before. 

I've left it a guessing game...  ;-)

darren