Subj : Re: Memory visibility and MS Interlocked instructions
To   : comp.programming.threads
From : Joe Seigh
Date : Fri Aug 26 2005 01:57 pm

Alexander Terekhov wrote:
> Joe Seigh wrote:
> 
>>Alexander Terekhov wrote:
>>
>>>To Joe: why don't you simply spend your weekend studying the archives.
>>>
>>>jupiter.robustserver.com/pipermail/cpp-threads_decadentplace.org.uk
>>>
>>
>>Done.
> 
> 
> Where can I buy that supersonic accelerator drug you're on?

There wasn't that much to catch up on.

> 
> 
>>1)  None of you know how to define semantics for atomic ops.
> 
> 
> Feel free to illuminate me.

It's not my job.  The burden of proof is on you since you've
taken it on yourselves to define this stuff.  You do have
the excuse that there's been no good examples of prior work
in this area, Posix having made a miserable failure of it
themselves.

> 
> 
>>2)  You have no examples, let alone compelling ones, on why
>>    anything more complex than simple memory ordering needs to be
>>    exposed at the api level rather than left as an implementation
>>    decision. 
> 
> 
> I'm just curious: what sort of complex implementation of simple memory 
> ordering do you have in your mind?

Wrong way around.  I asked you the question.  What usage example
requires api level semantics more comples than simple memory
ordering?  I already gave an example of the one exception I'm
aware of.

> 
> regards,
> alexander.


-- 
Joe Seigh

When you get lemons, you make lemonade.
When you get hardware, you make software.

.