Re: New file uploaded to boost

From: williamkempf_at_[hidden]
Date: 2001-05-22 16:52:38

--- In boost_at_y..., Beman Dawes <bdawes_at_a...> wrote:
> The whole thing with atomic_t might be a red herring.  It may well 
be that 
> thread_safe_int_least32_t needs a full set of integer operations.  
So even 
> if sizeof(int)==32, it can't just be a typedef for atomic_t.

I just realized what you're saying here.  I don't think there's any 
way to natively give a full set of integer operations to a type with 
out using a mutex.  I'm not aware of any platform that gives atomic 
multiplication, for instance.  You can't mix and match usage here so 
even if atomic addition is supported, since atomic multiplication 
isn't you'd have to use a mutex for both operations.  At that point 
the granularity of the lock combined with the usage would result in a 
very poor type indeed, and you'd likely find a better design would be 
to use a non-thread safe integer and a mutex at a higher level of 
granularity.  Atomic integer operations are really only useful for 
counting (increment and decrement) or swapping.

Bill Kempf