Re: check with SGML::Parser::OpenSP

* Ville Skyttä wrote:
>On a related note, I did some benchmarking with 0.7.0 on my lowlyish
>laptop yesterday, hitting a validation results page of a localhost URL
>that produces a couple of errors.  This was done with "ab", using a
>concurrency level of 5.  I don't have all the output available any more,
>but here are some order-of-magnitude results for served validation
>requests per second:
>
>Vanilla 0.7.0: about 0.58.

I get 0.70-0.75 for the CGI version regardless of whether 5 or 20
concurrent requests on my machine (latest version of everything).
I could not use a higher concurrency level as each perl process
uses about 17MB of memory and using more would mean it spends most
time swapping. check with my patch and some minor tweaks to make
it work with mod_perl (it does not like the main::* references)
and mod_perl 2.01 (with a minimal default configuration):

  Server Software:        Apache/2.0.54
  Server Hostname:        ...
  Server Port:            80
  
  Document Path:          ... check with local copy of msdn ...
  Document Length:        90014 bytes
  
  Concurrency Level:      50
  Time taken for tests:   390.281250 seconds
  Complete requests:      1000
  Failed requests:        962
     (Connect: 0, Length: 962, Exceptions: 0)
  Write errors:           0
  Total transferred:      16010642 bytes
  HTML transferred:       15775846 bytes
  Requests per second:    2.56 [#/sec] (mean)
  Time per request:       19514.063 [ms] (mean)
  Time per request:       390.281 [ms] (mean, across all concurrent requests)
  Transfer rate:          40.06 [Kbytes/sec] received
  
  Connection Times (ms)
                min  mean[+/-sd] median   max
  Connect:        0    5   9.7      0      93
  Processing:  2812 19099 4066.4  17766   36094
  Waiting:     2812 19092 4061.7  17765   36093
  Total:       2812 19104 4065.4  17781   36109
  
  Percentage of the requests served within a certain time (ms)
    50%  17781
    66%  18421
    75%  19375
    80%  20406
    90%  26734
    95%  27203
    98%  27609
    99%  29125
   100%  36109 (longest request)

With constant memory usage after the first x requests (with OpenSP
1.5.1 it probably would not constant as it has a huuuge memory leak).
Now, 2.56 is a bit better than 0.75...
-- 
Björn Höhrmann · mailto:bjoern@hoehrmann.de · http://bjoern.hoehrmann.de
Weinh. Str. 22 · Telefon: +49(0)621/4309674 · http://www.bjoernsworld.de
68309 Mannheim · PGP Pub. KeyID: 0xA4357E78 · http://www.websitedev.de/

Received on Friday, 12 August 2005 13:41:35 UTC