Concurrent collections

Simon Marlow marlowsd at gmail.com
Tue Apr 24 15:13:59 CEST 2012


On 20/04/2012 16:06, Johan Tibell wrote:

>     I am not sure it would be worth implementing a structure in Haskell
>     that is inherently concurrent. For me it was always enough to wrap
>     a persistent structure in an MVar, so updates are performed sequentially
>     and accesses can be made in parallel.
>
>
> I think it is worth implementing. We use the MVar strategy in the IO
> manager, but it scales poorly.

Are you referring to the priority queue here?  My vague recollection was 
that once we avoided the generational GC issue and fixed the blackhole 
implementation, there weren't any remaining problems.

I'd really like to see some applications that are suffering here, and 
analyse whether it really is contention for the shared data structure, 
and if so, whether there's anything else we can do.  Ryan, do you have 
anything?

Cheers,
	Simon




>
>     Such a structure would probably have to be strict and in an ST monad,
>     to get a usable and deterministic behaviour.
>
>     Any thoughts?
>
>     Cheers,
>     Milan
>
>     _______________________________________________
>     Libraries mailing list
>     Libraries at haskell.org <javascript:;>
>     http://www.haskell.org/mailman/listinfo/libraries
>
>
>
> _______________________________________________
> Libraries mailing list
> Libraries at haskell.org
> http://www.haskell.org/mailman/listinfo/libraries




More information about the Libraries mailing list