Re: [Algorithms] General purpose task parallel threading approach

[Nicholas \Indy\ R. <ar...@gm...>]

On Tue, Apr 14, 2009 at 3:26 PM,  <asy...@gm...> wrote:
> That's not really true, you can still execute other tasks the same way in a
> recursive manner (but keep away of dependencies on something up the worker
> thread's stack).

I'm sorry, that statement makes little sense to me.

> Besides, even if it would be fully true, then "the vast majority of the
> functionality" is actually a ~2k of code which you never execute, is this a
> "generalization" overhead you can't afford?

~2k loc have a tendency to muck up performance, and other such things.
But alas, when I have to include 2k useless loc for functionality
better provided by a specific library that takes only a few hours to
write/debug/test I'd always go for the later, so yes, that is
*certainly* a generalization overhead I cannot afford.

Indy