[Bf-committers] Re: Thoughts on speeding up...

[Jean-Luc Peuriere]

Le 24 mars 05, à 22:27, GSR - FR a écrit :

> All the Intels with Hyper Toasting^WThreading could * make use of
> that, plus the new plans for dual core (and of course the combination
> of all them: SMT, discrete and multicore SMP) seem to "paint" a future
> where speed ups have to be won by paralelizing. So next times they
> will be a bit more common.
>

Hmm.  In this case Hyper Toasting is appropriate, as with a single 
floating
point unit, both threads will compete for it and often result in not 
better,
or even worse rendering times. Dont trust Intel Marketers ;)

HT can be very useful in some cases, but here it is really not a 
solution,
only real SMP will get true benefits.

As for supporting more than 2 cores, yes that can be fine, but the 
logical
  solution for render is to launch more than one instance, each rendering
at its own path its own frame. No scheduling problem this way nor coding
complexity.

Unless cell processors makes it in the personal computer area (which 
for the
moment is still science-fiction, while  we should see them soon in many 
appliances),
2 or at most 4 rendering threads seems fine for the time being.

Now, threading long tasks allow to return to UI handling sooner and  
can make
the application more responsive, but that is a completely different 
problem, and
not easy

-- 
Jean-Luc