[rabbitmq-discuss] Simple benchmark and results

Fri Oct 23 09:49:57 BST 2009

On Fri, Oct 23, 2009 at 4:04 AM, David Glaubman <dglaubman at acm.org> wrote:
>
> Not a showstopper, since with the partial fix, Rabbit should be sufficient to
> show feasibility of the approach.
>
> Still, sooner is better than later, and I hope you folks will take a close
> look at network performance on .Net/Windows, since it seems to differ from
> Mono (or even Java on Windows).

Understood.

> I also think it may be worthwhile to consider programming against Socket
> interface rather than TCPClient to maximize performance.

Is that something you could have a crack at?

alexis

> (just my 0.01334 Euros.)
>
> David
>
>
>
> Matthias Radestock-2 wrote:
>>
>> David,
>>
>> David Glaubman wrote:
>>> Buffering the int/short Writes in a modified version of
>>> NetworkBinaryWriter,
>>> I got around 7.5K messages per sec.:  (5X speedup)
>>
>> The lack of buffering in the .net client is an issue we noticed a while
>> ago and we filed a bug to look into it. Our results at the time didn't
>> show as dramatic a difference as your tests. I've increased the severity
>> of that bug now, so we'll address it sooner.
>>
>> Is this a showstopper for you or is the performance as it stands
>> sufficient for your intended use case?
>>
>> Thanks for your help in tracking this down.
>>
>>
>> Regards,
>>
>> Matthias.
>>
>> _______________________________________________
>> rabbitmq-discuss mailing list
>> rabbitmq-discuss at lists.rabbitmq.com
>> http://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss
>>
>>
>
> --
> View this message in context: http://www.nabble.com/Simple-benchmark-and-results-tp25968960p26020282.html
> Sent from the RabbitMQ mailing list archive at Nabble.com.
>
>
> _______________________________________________
> rabbitmq-discuss mailing list
> rabbitmq-discuss at lists.rabbitmq.com
> http://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss
>