Discussion:
Is there some theoretical reason hash tables cannot be implemented with a "mutable spine" in Haskell?
(too old to reply)
Casey Hawthorne
2009-04-13 18:18:18 UTC
Permalink
That would be a mutable array.
--
Regards,
Casey
Dan Doel
2009-04-13 19:42:25 UTC
Permalink
Post by Casey Hawthorne
That would be a mutable array.
That is how they are implemented.

-- Dan
Jon Harrop
2009-04-18 19:03:17 UTC
Permalink
Post by Dan Doel
Post by Casey Hawthorne
That would be a mutable array.
That is how they are implemented.
The problem I highlighted with the current hash table implementation
provided with GHC is that the library code is fine but a perf bug in the
GHC garbage collector makes mutating arrays of boxed values asymptotically
slower than it should be.
--
Dr Jon D Harrop, Flying Frog Consultancy Ltd.
http://www.ffconsultancy.com/?u
Loading...