ConcurrentHashMap

View: New views
2 Messages — Rating Filter:   Alert me  

ConcurrentHashMap

by Ohad Shacham :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Hi,



I noticed that in version 6.0.20 several Hashtables and HashMaps were
replaced by ConcurrentHashMaps. I have noticed that many Hashtables were not
replaced; therefore, I assume that a contention detection tool was used to
identify the contended Hashtables. I would be happy to know if this is the
case and also whether a scalability and/or performance improvement was
achieved by switching these collections with concurrent ones.



Thanks,

Ohad

Parent Message unknown Re: ConcurrentHashMap

by markt-2 :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

> From: Ohad Shacham <ohad.shacham@...>
> To: dev@...
> Sent: 08/11/09, 14:47:25
> Subject: ConcurrentHashMap
>
> Hi,
>
>
>
> I noticed that in version 6.0.20 several Hashtables and HashMaps were
> replaced by ConcurrentHashMaps. I have noticed that many Hashtables were not
> replaced; therefore, I assume that a contention detection tool was used to
> identify the contended Hashtables.
Not to my knowledge.

> I would be happy to know if this is the
> case and also whether a scalability and/or performance improvement was
> achieved by switching these collections with concurrent ones.
Based on general experience I'd say yes, mainly for multi-core machines. In specific cases I don't recall any measurements or results.

The corresponding comit messages should be the best source of information for individual changes.

Mark


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@...
For additional commands, e-mail: dev-help@...