Subj : Re: Challenge: Multithreading & Synchronization
To   : comp.programming.threads
From : David Schwartz
Date : Thu May 26 2005 12:10 pm


"Peter Dimov" <pdimov@gmail.com> wrote in message 
news:d74kp4$85u$1@domitilla.aioe.org...

> I'm relatively new to this group, so if it is customary here to have 
> "atomic
> reads and writes" mean "atomic reads and writes with respect to the CPU
> instruction stream", apologies. It's hard to see how this definition of
> "atomic" could be useful or even sensible, but to each his own.

    Sadly, the word "atomic" is used several different ways. I try to remind 
people that it is often used to mean "a single instruction that cannot be 
interrupted until complete" so that they don't make the mistake of thinking 
that normal atomic types and functions are suitable for multi-threaded 
programming on platforms that can support multiple logical or physical CPUs.

    DS

.