[ofiwg] allowing aliasing in fetch atomics?
Dave Goodell (dgoodell)
dgoodell at cisco.com
Wed Nov 11 08:02:00 PST 2015
On Nov 11, 2015, at 9:47 AM, Jeff Hammond <jeff.science at gmail.com> wrote:
> Is there a reason why buf and result can't be restrict pointers? I see zero value in enabling the case of aliased local buffers when the side effects are (1) a const buffer is modified and (2) the potential code optimization demonstrated in my pull request is disallowed (without specializing the macro for the aliased and non-aliased cases, of course).
Let's flip that around: do you have any empirical evidence that this API design decision is actually affecting performance in a noticeable way? "I would like it to look differently, and I *think* there's optimization potential" is a weak argument.
More information about the ofiwg