You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@mina.apache.org by Dong Tang <dt...@gmail.com> on 2008/08/04 04:22:39 UTC

Re: I noticed your async io

re-send.

On 8/3/08, Dong Tang <dt...@gmail.com> wrote:
>
> Hi, Mina Owner
>
> I noticed your Asynch IO,
>
> Do you have any perforance test result compare to regular io or NIO?
>
> Thnx. David
>

Re: I noticed your async io

Posted by Dong Tang <dt...@gmail.com>.

thnx for your comments.

On 8/4/08, Mike Heath <mh...@apache.org> wrote:
>
> Dong Tang wrote:
>
>> re-send.
>>
>> On 8/3/08, Dong Tang <dt...@gmail.com> wrote:
>>
>>
>>> Hi, Mina Owner
>>>
>>> I noticed your Asynch IO,
>>>
>>> Do you have any perforance test result compare to regular io or NIO?
>>>
>>>
>> Using a thread pool with FileChannels is faster than POSIX AIO on Linux.
>  There are different techniques that can be used on Linux that should
> improve performance but I haven't experimented with that at all.
>
> I think you would be better served using NIO2.  You can get snapshot
> binaries of OpenJDK with NIO2 support here:
> http://download.java.net/jdk7/jsr203/binaries/
>
> -Mike
>

Re: I noticed your async io

Posted by Mike Heath <mh...@apache.org>.

Dong Tang wrote:
> re-send.
>
> On 8/3/08, Dong Tang <dt...@gmail.com> wrote:
>   
>> Hi, Mina Owner
>>
>> I noticed your Asynch IO,
>>
>> Do you have any perforance test result compare to regular io or NIO?
>>     
Using a thread pool with FileChannels is faster than POSIX AIO on 
Linux.  There are different techniques that can be used on Linux that 
should improve performance but I haven't experimented with that at all.

I think you would be better served using NIO2.  You can get snapshot 
binaries of OpenJDK with NIO2 support here: 
http://download.java.net/jdk7/jsr203/binaries/

-Mike