is_pmc_ptr

Andrew Whitworth wknight8111 at gmail.com
Sun Sep 4 12:31:24 UTC 2011


I've prototyped this idea in the whiteknight/pmc_is_ptr branch. It's
passing tests in the branch, but I haven't done any benchmarking yet.
Code review, additional testing, and benchmark results would all be
appreciated

--Andrew Whitworth



On Sun, Sep 4, 2011 at 5:18 AM, Nicholas Clark <nick at ccl4.org> wrote:
> On Sun, Sep 04, 2011 at 01:06:47AM +0200, Jonathan Worthington wrote:
>
>> One obvious fix would be to just keep an extent list: an array of
>> (start,end) pairs, hanging off the pool header, which we can very
>> cheaply scan through (since it's just a bunch of values laid out
>> contiguously in memory). That way, we don't go chasing all over RAM, and
>
> I think you can binary search it, as it's effectively an inversion list.
> O(log n) will beat O(n) for any non-trivial size :-)
>
>
> Nicholas Clark
> _______________________________________________
> http://lists.parrot.org/mailman/listinfo/parrot-dev
>


More information about the parrot-dev mailing list