[swift-dev] Performance issues in automatic reference counting (ARC)?

Brian Gesiak modocache at gmail.com
Sat Dec 17 14:10:57 CST 2016


Great, thanks! That article was exactly what I had in mind. :)

- Brian Gesiak


On Sat, Dec 17, 2016 at 2:47 PM, Mark Lacey <mark.lacey at apple.com> wrote:

>
> On Dec 17, 2016, at 11:13 AM, Brian Gesiak via swift-dev <
> swift-dev at swift.org> wrote:
>
> Hello all!
>
> I really enjoyed Chris Lattner's slides from his talk at IBM <
> http://researcher.watson.ibm.com/researcher/files/us-lmandel/lattner.pdf
> >.
>
> The speaker notes mention ARC:
>
> "There are two principle downsides to ARC that people cite: one is the
> need for atomic increment/decrements, which can be slow." [...] "The
> performance problems it can cause are real in some important cases"
>
> Can someone point me to a good resource that explains these problems?
>
>
> This might be a good starter post: https://fgiesen.
> wordpress.com/2014/08/18/atomics-and-contention
>
> I guess atomic reference count changes create overhead in multithreaded
> applications?
>
>
> There is overhead even in single-threaded applications, as the post above
> discusses. The overhead depends very much on the details of the particular
> microarchitecture.
>
> Are there more detailed explorations into this topic?
>
>
> If you’re looking for something Swift- or ARC-specific I don’t know of
> anything.
>
> Mark
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.swift.org/pipermail/swift-dev/attachments/20161217/0f22961a/attachment.html>


More information about the swift-dev mailing list