RE: [boost] Re: Threads: New Win32 Implementation

From: Bill Wade (bill.wade_at_[hidden])
Date: 2000-09-07 13:24:03

> From: William Kempf [mailto:sirwillard_at_[hidden]]
>
> A corner case indeed!  It's very rarely a good idea to use node level
> locking here.  The granularity is much too small and will lead to
> very poor performance.

In practice you'd lock enough nodes at once (or make node processing
expensive enough) that the amortized cost of the locks was less than the
advantage gained by allowing multiple processors to attack the problem.

If you have just one processor and are using MT because it is a good way to
express the solution to the problem, it becomes much more likely that
locking the entire list will be appropriate.