Eugen Leitl wrote:
On Tue, Apr 03, 2012 at 08:19:53AM -0700, David Barbour wrote:
That said, I also disagree with Tom, there: design complexity doesn't need
to increase with parallelism. The tradeoff between complexity vs.
parallelism is more an artifact of sticking with imperative programming.
It's not just imperative programming. The superficial mode of human
cognition is sequential. This is the problem with all of mathematics
and computer science as well.
If you look at MPI debuggers, it puts people into a whole other
universe of pain that just multithreading.
Sort of depends on how you architect things. Email is one huge parallel
system. Nobody even thinks of debugging the global flow of email. It's
a probabilistic approach.
Dataflows and pipelines can be parallelized without issue and remain
deterministic. If we push more of the parallelism to the code, the hardware
can also be less complex - i.e. less architecture to hide latency for
memory access.
Global memory doesn't scale in a relativistic universe. Ditto cache
coherence for already reasonable small number of caches.
So, we don't really have a choice other to stop worrying, and learn
to love parallelism.
True that. :-)
--
In theory, there is no difference between theory and practice.
In practice, there is. .... Yogi Berra
_______________________________________________
fonc mailing list
fonc@vpri.org
http://vpri.org/mailman/listinfo/fonc
