Re: Data.HashTable performance

Actually, Data.HashTable got some good attention recently from Simon Marlow (modifying the GC), and Johan Tibell and others looking at containers performance in general: http://stackoverflow.com/questions/3058529/curious-about-the-hashtable-probl... http://blog.johantibell.com/2011/03/video-of-my-hashing-based-containers.htm... Certainly, being in IO limits the utility of a hashtable library (since it isn't thread safe or persistent by default), but that doesn't mean the structure gets neglected. Johan, what's your thoughts on the status? -- Don On Tue, Apr 5, 2011 at 8:31 PM, Thomas DuBuisson wrote:

The hashtable in base is used very little by the Haskell community, partly due to it being IO heavy and partly a result of its known performance issues that no one seems motivated to address.  I suggest you look at the unordered-containers package which is rather new but positioned to become the de facto standard.  Currently, many people needing a key/value mapping lean on Data.Map from the containers package (a balanced tree).

Cheers, Thomas

On Tue, Apr 5, 2011 at 5:53 PM, Frank Kuehnel wrote:

...

Out of curiosity, I've tested a few Hashtable implementations: http://zufaellige-reflektion.blogspot.com/2011/04/associative-arrays-showdow... Are there any clues why Haskell's Data.HashTable doesn't perform better? I would have expected a similar performance like with F# on the Mono 2.0 platform. Thanks, Frank

_______________________________________________ Libraries mailing list Libraries@haskell.org http://www.haskell.org/mailman/listinfo/libraries

_______________________________________________ Libraries mailing list Libraries@haskell.org http://www.haskell.org/mailman/listinfo/libraries