[cpp-threads] A question about N2153

[Raul Silvera]

As per the definitions in N2153, an ordered RMW has the acquire and release
orderings apply to both the load and store implied by the operation.

--
Raúl E. Silvera         IBM Toronto Lab   Team Lead, Toronto Portable
Optimizer (TPO)
Tel: 905-413-4188 T/L: 969-4188           Fax: 905-413-4854
D2/KC9/8200/MKM



                                                                           
             "Peter Dimov"                                                 
             <pdimov at mmltd.net                                             
             >                                                          To 
             Sent by:                  "C++ threads standardisation"       
             cpp-threads-bounc         <cpp-threads at decadentplace.org.uk>  
             es at decadentplace.                                          cc 
             org.uk                                                        
                                                                   Subject 
                                       Re: [cpp-threads] A question about  
             01/19/07 10:35 AM         N2153                               
                                                                           
                                                                           
             Please respond to                                             
                C++ threads                                                
              standardisation                                              
             <cpp-threads at deca                                             
             dentplace.org.uk>                                             
                                                                           
                                                                           




Raul Silvera wrote:

> N2153 is trying to propose a model that does not overtly favor any
> specific niche CPU. We're looking for a trade-off between
> programmability and performance across the whole spectrum of parallel
> architectures.

One more question, can you clarify the semantics of RMW_ordered? Is it
equivalent to RMW_acquire_release, where the acquire applies to the read
part, and the release applies to the write part, or is it ordered more
strongly (total systemwide order on all RMW_ordered, useful for building SC

atomics)?


--
cpp-threads mailing list
cpp-threads at decadentplace.org.uk
http://www.decadentplace.org.uk/cgi-bin/mailman/listinfo/cpp-threads