hc-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ortwin Glück <ortwin.glu...@nose.ch>
Subject Re: DO NOT REPLY [Bug 30388] - Performance Increase by setting the initial ByteArrayOS length
Date Wed, 11 Aug 2004 06:54:04 GMT

bugzilla@apache.org wrote:

> ------- Additional Comments From ghexsel@uol.com.br  2004-08-10 17:12 -------
> I have a complete template system that browses and parses webpages using
> multi-line regex'es.  The test page was about 6-10K, on a local server with
> little delay.  The measure was on the number of pages I could process in the
> whole system with a limit on memory (256MB).  I can't send you the complete code
> (my boss would hang me), but what I can say is that before that method was one
> of the main bottlenecks, now it's the regular expression code.  The system being
> tested is massively parallel, limited only by the number of threads on the
> MultiThreadedHttpConnectionManager and on my executor (I'm using Doug Lea's
> excelent concurrent library to manage works).  My settings were 500-1000
> concurrent conns to that website.  Anything else?

When dealing with such a complex system, just measuring the overall 
throughput can be totally wrong and misleading. I suggest you run a 
decent profiling of your application to see where the *real* bottlenecks 
are. I mean 1000 concurrent connections to a webserver are extreme.

  NOSE applied intelligence ag

  ortwin glück                      [www]      http://www.nose.ch
  software engineer
  hardturmstrasse 171               [pgp id]           0x81CF3416
  8005 zürich                       [office]      +41-1-277 57 35
  switzerland                       [fax]         +41-1-277 57 12

To unsubscribe, e-mail: commons-httpclient-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-httpclient-dev-help@jakarta.apache.org

View raw message