Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.programming.threads > #1777

Re: forcing the compiler to reload from memory with c++0x

Show all headers | View raw

David Jobet <david.jobet@free.fr> writes:

> 1.10/25 seems to be talking about a store right ?

Yes.

> So, if I'm doing atomic_int.load(std::memory_order_relaxed)
>
> does 1.10/25 still apply ?

Your load must see one of the values written that is consistent with any
other memory ordering constraints that apply.

> If it applies, then it means that an atomic load relaxed implies :
> - no ordering guarantee (that one everybody agrees)
> - the load cannot be cached/optimized out, because regardless of the 
> ordering model used, we're talking about an atomic here, and operation 
> on atomics (even load) must become visible in a finite time (I wasn't 
> aware of that one)

Repeated loads of the same variable can be optimized out unless
intervening operations impose ordering constraints. However, it is the
"early" loads that can be optimized out, not the "later" ones: the
compiler cannot read the value once and then cache it forever, as this
means that stores done by other threads won't become visible to that
thread, and 1.10/25 is violated. The compiler can skip any finite number
of loads, but must sooner or later execute a real load.

> Corrolary : if I'm doing my own impl of atomic load relaxed in c++98, 
> does it mean that even a load relaxed implies a memory clobber so that 
> the compiler cannot optimize the value out ?

With gcc, then yes.

Anthony
-- 
Author of C++ Concurrency in Action     http://www.stdthread.co.uk/book/
just::thread C++0x thread library             http://www.stdthread.co.uk
Just Software Solutions Ltd       http://www.justsoftwaresolutions.co.uk
15 Carrallack Mews, St Just, Cornwall, TR19 7UL, UK. Company No. 5478976

Back to comp.programming.threads | Previous | Next | Find similar

Thread

Re: forcing the compiler to reload from memory with c++0x Anthony Williams <anthony.ajw@gmail.com> - 2011-01-31 09:11 +0000

csiph-web