Mailing-List: contact dev-help@mina.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@mina.apache.org
Received-SPF: pass (herse.apache.org: local policy)
Message-ID: <A0933DFE53EB44E4B11D0C0382887A07@UberQuad>
From: "Michael Grundvig" <mike@electrotank.com>
To: <dev@mina.apache.org>
References: <4627795D.9070107@webcogs.com>
 <9166C436-C4AB-4147-B372-720EEE9E5057@apache.org>
 <46279329.4090902@webcogs.com>
 <9f066ee90704190917w4a5102b6w34185fc3589f19df@mail.gmail.com>
 <46279C27.7080103@webcogs.com>
 <45a054f50704190949y1368d67bx5ee3079ac38ecfee@mail.gmail.com>
 <4627A61C.8010803@webcogs.com>
 <768dcb2e0704191708l679691dbt79ce2cfd17d18937@mail.gmail.com>
 <768dcb2e0704191715k1b84b2c5g4f6d14eac47af853@mail.gmail.com>
 <462A1879.708@webcogs.com> <C74E8639-044E-41EE-9B16-E8B397F21984@mac.com>
In-Reply-To: <C74E8639-044E-41EE-9B16-E8B397F21984@mac.com>
Subject: Re: 1000+ simultaneous connections with data transfer?
Date: Sat, 21 Apr 2007 11:16:13 -0500
MIME-Version: 1.0
Content-Type: text/plain;
	format=flowed;
	charset="iso-8859-1";
	reply-type=response
Content-Transfer-Encoding: 7bit

I'm using Trove for a large project currently. It's faster at times and 
slower at others with the newest JDK. Overall though, if you can get away 
with a primitive as the key, it works quite well. The licensing is a problem 
though as it's GPL.

Mike

----- Original Message ----- 
From: "Ming Fang" <mingfang@mac.com>
To: <dev@mina.apache.org>
Sent: Saturday, April 21, 2007 11:09 AM
Subject: Re: 1000+ simultaneous connections with data transfer?


> Has anyone try using the Trove HashMap implementation(http:// 
> trove4j.sourceforge.net/)?
> For some tests it is many times faster than the JDK one.
>
> On Apr 21, 2007, at 9:58 AM, Richard Lowe wrote:
>
>>   I found that HashMap lookups with this amount of packets can be 
>> particularly expensive, as can creating byte arrays (even tiny  ones).  I 
>> find that Mina session attributes are very useful.  I  know all of that 
>> sounds obvious, but you'd be surprised how simple  little operations can 
>> stack up.
>>
>> If you have to write a byte array to the IoHandler adapter, use 
>> ByteBuffer.wrap(b) - this seems pretty fast.
>>
>> I hope that might be useful for anybody else trying to do anything 
>> similar!
>>
>> Cheers,
>> Richard.
>> --
>>
>> Trustin Lee wrote:
>>> Additionally, did you try to turn on 'tcp no delay' option? (this
>>> doesn't affect CPU utilization but might improve throughput)
>>>
>>> ((SocketSessionConfig) cfg.getSessionConfig()).setTcpNoDelay(true);
>>>
>>> and... It would be nice to see your IoHandler implementation code.
>>>
>>> HTH,
>>> Trustin
>>
>