http://bugzilla.gdcproject.org/show_bug.cgi?id=84
--- Comment #6 from Johannes Pfau <johannesp...@gmail.com> 2014-02-13 15:01:52
GMT ---
Not into registers might be OK, but if we have code like this:
-------------
shared int a;
void doSomething()
{
for(int i = 0; i < 10; i++)
atomicStore(a, i);
}
-------------
I think it is legal* to optimize the for loop into a single assignment, 'a =
9', for shared variables. This wouldn't be allowed for volatile variables.
* http://www.drdobbs.com/parallel/volatile-vs-volatile/212701484
Basically the 'as if' rule: If the 'doSomething' thread executes much faster
than a thread reading 'a' the reading thread might only see the final value, 9,
anyway.
--
Configure bugmail: http://bugzilla.gdcproject.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.
