Page 4 of 6 FirstFirst 123456 LastLast
Results 76 to 100 of 143
  1. #76
    Setup port 8081 as 1-CPU, port 8082 as 2-CPU, port 8084 as 4-cpu, 8088 as 8-CPU

    4 ab clients against 4-CPU result:

    5824.48+5775.94+5704.83+5646.85=22952.1


    Code:
    
    [root@centos-magic conf]# ab -n 100000 -c 100 http://localhost:8084/test.html & ab -n 100000 -c 100 http://localhost:8084/test.html & ab -n 100000 -c 100 http://localhost:8084/test.html & ab -n 100000 -c 100 http://localhost:8084/test.html &
    [1] 15923
    [2] 15924
    [3] 15925
    [4] 15926
    [root@centos-magic conf]# This is ApacheBench, Version 2.0.40-dev <$Revision: 1.146 $> apache-2.0
    Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
    Copyright 2006 The Apache Software Foundation, http://www.apache.org/
    This is ApacheBench, Version 2.0.40-dev <$Revision: 1.146 $> apache-2.0
    Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
    Copyright 2006 The Apache Software Foundation, http://www.apache.org/
    
    Benchmarking localhost (be patient)
    
    Benchmarking localhost (be patient)
    This is ApacheBench, Version 2.0.40-dev <$Revision: 1.146 $> apache-2.0
    Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
    Copyright 2006 The Apache Software Foundation, http://www.apache.org/
    
    Benchmarking localhost (be patient)
    This is ApacheBench, Version 2.0.40-dev <$Revision: 1.146 $> apache-2.0
    Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
    Copyright 2006 The Apache Software Foundation, http://www.apache.org/
    
    Benchmarking localhost (be patient)
    Completed 10000 requests
    Completed 10000 requests
    Completed 10000 requests
    Completed 10000 requests
    Completed 20000 requests
    Completed 20000 requests
    Completed 20000 requests
    Completed 20000 requests
    Completed 30000 requests
    Completed 30000 requests
    Completed 30000 requests
    Completed 30000 requests
    Completed 40000 requests
    Completed 40000 requests
    Completed 40000 requests
    Completed 40000 requests
    Completed 50000 requests
    Completed 50000 requests
    Completed 50000 requests
    Completed 50000 requests
    Completed 60000 requests
    Completed 60000 requests
    Completed 60000 requests
    Completed 60000 requests
    Completed 70000 requests
    Completed 70000 requests
    Completed 70000 requests
    Completed 70000 requests
    Completed 80000 requests
    Completed 80000 requests
    Completed 80000 requests
    Completed 80000 requests
    Completed 90000 requests
    Completed 90000 requests
    Completed 90000 requests
    Completed 90000 requests
    Finished 100000 requests
    
    
    Server Software:        LiteSpeed
    Server Hostname:        localhost
    Server Port:            8084
    
    Document Path:          /test.html
    Document Length:        101 bytes
    
    Concurrency Level:      100
    Time taken for tests:   17.168913 seconds
    Complete requests:      100000
    Failed requests:        0
    Write errors:           0
    Total transferred:      33705392 bytes
    HTML transferred:       10101616 bytes
    Requests per second:    5824.48 [#/sec] (mean)
    Time per request:       17.169 [ms] (mean)
    Time per request:       0.172 [ms] (mean, across all concurrent requests)
    Transfer rate:          1917.13 [Kbytes/sec] received
    
    Connection Times (ms)
                  min  mean[+/-sd] median   max
    Connect:        0    2   2.5      2      42
    Processing:     3   14   5.8     12      50
    Waiting:        0    6   3.5      5      32
    Total:          4   16   7.1     14      72
    
    Percentage of the requests served within a certain time (ms)
      50%     14
      66%     15
      75%     16
      80%     18
      90%     28
      95%     34
      98%     38
      99%     42
     100%     72 (longest request)
    Finished 100000 requests
    
    
    Server Software:        LiteSpeed
    Server Hostname:        localhost
    Server Port:            8084
    
    Document Path:          /test.html
    Document Length:        101 bytes
    
    Concurrency Level:      100
    Time taken for tests:   17.313188 seconds
    Complete requests:      100000
    Failed requests:        0
    Write errors:           0
    Total transferred:      33702359 bytes
    HTML transferred:       10100707 bytes
    Requests per second:    5775.94 [#/sec] (mean)
    Time per request:       17.313 [ms] (mean)
    Time per request:       0.173 [ms] (mean, across all concurrent requests)
    Transfer rate:          1900.98 [Kbytes/sec] received
    
    Connection Times (ms)
                  min  mean[+/-sd] median   max
    Connect:        0    2   2.3      2      34
    Processing:     3   14   6.0     12      49
    Waiting:        0    6   3.6      5      36
    Total:          4   16   7.2     14      55
    
    Percentage of the requests served within a certain time (ms)
      50%     14
      66%     15
      75%     16
      80%     18
      90%     29
      95%     34
      98%     38
      99%     41
     100%     55 (longest request)
    Finished 100000 requests
    
    
    Server Software:        LiteSpeed
    Server Hostname:        localhost
    Server Port:            8084
    
    Document Path:          /test.html
    Document Length:        101 bytes
    
    Concurrency Level:      100
    Time taken for tests:   17.529004 seconds
    Complete requests:      100000
    Failed requests:        0
    Write errors:           0
    Total transferred:      33704718 bytes
    HTML transferred:       10101414 bytes
    Requests per second:    5704.83 [#/sec] (mean)
    Time per request:       17.529 [ms] (mean)
    Time per request:       0.175 [ms] (mean, across all concurrent requests)
    Transfer rate:          1877.69 [Kbytes/sec] received
    
    Connection Times (ms)
                  min  mean[+/-sd] median   max
    Connect:        0    2   2.6      2      37
    Processing:     4   14   6.0     12      54
    Waiting:        0    6   3.6      5      42
    Total:          5   16   7.4     14      68
    
    Percentage of the requests served within a certain time (ms)
      50%     14
      66%     15
      75%     17
      80%     19
      90%     29
      95%     34
      98%     39
      99%     41
     100%     68 (longest request)
    Finished 100000 requests
    
    
    Server Software:        LiteSpeed
    Server Hostname:        localhost
    Server Port:            8084
    
    Document Path:          /test.html
    Document Length:        101 bytes
    
    Concurrency Level:      100
    Time taken for tests:   17.708981 seconds
    Complete requests:      100000
    Failed requests:        0
    Write errors:           0
    Total transferred:      33705055 bytes
    HTML transferred:       10101515 bytes
    Requests per second:    5646.85 [#/sec] (mean)
    Time per request:       17.709 [ms] (mean)
    Time per request:       0.177 [ms] (mean, across all concurrent requests)
    Transfer rate:          1858.66 [Kbytes/sec] received
    
    Connection Times (ms)
                  min  mean[+/-sd] median   max
    Connect:        0    2   2.4      2      27
    Processing:     4   14   6.3     12      63
    Waiting:        0    6   3.7      5      45
    Total:          4   17   7.4     14      71
    
    Percentage of the requests served within a certain time (ms)
      50%     14
      66%     16
      75%     17
      80%     21
      90%     30
      95%     34
      98%     38
      99%     41
     100%     71 (longest request)
    
    [1]   Done                    ab -n 100000 -c 100 http://localhost:8084/test.html
    [2]   Done                    ab -n 100000 -c 100 http://localhost:8084/test.html
    [3]-  Done                    ab -n 100000 -c 100 http://localhost:8084/test.html
    [4]+  Done                    ab -n 100000 -c 100 http://localhost:8084/test.html
    [root@centos-magic conf]#    
    
    LiteSpeed Web Acceleration Platform by https://www.litespeedtech.com
    Apache drop-in replacement. Triple server capacity with 10X performance increases.
    Ultimate web serving platform for WordPress, Magento and other web applications.
    Turbo charging all WordPress sites hosted on your server with a single click!

  2. #77
    Join Date
    Jan 2005
    Posts
    2,203
    Seems nwmcsween's numbers are higher than mistwang's? Where do you get 10580 from?

  3. #78
    Join Date
    Mar 2009
    Location
    /usr/bin/perl
    Posts
    971
    Quote Originally Posted by HD Fanatic View Post
    Seems nwmcsween's numbers are higher than mistwang's? Where do you get 10580 from?
    He uploaded his own customized binaries, apparently.
    Ask me about CloudCentrum (coming soon) -- The complete, turn-key cloud software solution

  4. #79
    Quote Originally Posted by HD Fanatic View Post
    Seems nwmcsween's numbers are higher than mistwang's? Where do you get 10580 from?
    10580 was for four runs at the same time

  5. #80
    Join Date
    Jan 2005
    Posts
    2,203
    I see, so Litespeed still outperforms nginx by a huge margin.

  6. #81
    Quote Originally Posted by HD Fanatic View Post
    I see, so Litespeed still outperforms nginx by a huge margin.
    nginx beat litespeed on a single ab instance serving 8000 rps, litespeed beat nginx on 4 ab instances serving 20000 rps. This is probably due to fine grain locking in litespeed?
    Last edited by nwmcsween; 01-23-2010 at 06:24 AM.

  7. #82
    LiteSpeed can deliver around 7700 req/s with around 30% cpu utilization on single CPU, you can test it against port 8081. How much CPU does nginx uses to serving around 7900 req/s?
    LiteSpeed Web Acceleration Platform by https://www.litespeedtech.com
    Apache drop-in replacement. Triple server capacity with 10X performance increases.
    Ultimate web serving platform for WordPress, Magento and other web applications.
    Turbo charging all WordPress sites hosted on your server with a single click!

  8. #83
    Let's moving forward with pending benchmark tests.

    Lighttpd: small file keepalive test
    Nginix: small file keepalive test, PHP hello world test

    Lighttpd only use one worker process, it can be configured to use multiple workers, you can give a try to get better result.
    LiteSpeed Web Acceleration Platform by https://www.litespeedtech.com
    Apache drop-in replacement. Triple server capacity with 10X performance increases.
    Ultimate web serving platform for WordPress, Magento and other web applications.
    Turbo charging all WordPress sites hosted on your server with a single click!

  9. #84
    Join Date
    Jun 2003
    Location
    London, UK
    Posts
    1,765
    Could we try something "real world" like hitting a base wordpress install with the default theme?
    Darren Lingham - Stablepoint Hosting
    Stablepoint - Cloud Web Hosting without compromise
    We provide industry-leading cPanel™ web hosting in 80+ global cities.

  10. #85
    Join Date
    Oct 2009
    Posts
    865
    It would certainly be a lot more relevant with some real-life numbers containing actual processing and database requests. While handling tons and tons of small requests fast could be useful during a DDoS, you can get a lot of hardware firewall for $800/month.

  11. #86
    Join Date
    Mar 2009
    Location
    /usr/bin/perl
    Posts
    971
    Quote Originally Posted by dazmanultra View Post
    Could we try something "real world" like hitting a base wordpress install with the default theme?
    Yes, what about siege instead of ab? Let's give that a try.
    Ask me about CloudCentrum (coming soon) -- The complete, turn-key cloud software solution

  12. #87
    Join Date
    Mar 2008
    Posts
    1,717
    Quote Originally Posted by dazmanultra View Post
    Could we try something "real world" like hitting a base wordpress install with the default theme?
    I don't think the types of files being served are skewing the results any (and I'm an Apache fanboy, so I'd be the first to cry foul!) - any wordpress site that's going to be taking 20+k hits/sec is going to be using super cache, and I doubt (though I admit I'm not terribly well educated on the subject) the difference between these tiny files and the average static files super cache builds is going to do much of anything except affect network performance.
    I used to run the oldest commercial Mumble host.

  13. #88
    Join Date
    Nov 2005
    Posts
    3,944
    These results look strange to me. Personally I have tested with NGINX on my single processor quad-core and have gotten results equal to or greater than what is shown and for a more intense HTML file (>1KB) and much better results for A hello world PHP.

    I think there needs to be more organization with your testing as results are sporadic as of now.


  14. #89
    Join Date
    Aug 2004
    Location
    Canada
    Posts
    3,785
    Quote Originally Posted by dazmanultra View Post
    Could we try something "real world" like hitting a base wordpress install with the default theme?

    Sort of silly because then you're testing things outside the web servers control that will not change anything. The speed at which the wordpress portion of the processing is done will not change between servers. What changes even in a PHP test is before anything else happens as far as PHP. So a hello world test makes far more sense. All you're doing with the wordpress test is slowing each one down equally. They cannot magically make PHP itself faster. They just make the SAPI portion faster nothing else.
    Tony B. - Chief Executive Officer
    Hawk Host Inc. Proudly serving websites since 2004
    Quality Shared and Cloud Hosting
    PHP 5.2.x - PHP 8.1.X Support!

  15. #90
    Join Date
    Mar 2009
    Location
    /usr/bin/perl
    Posts
    971
    Quote Originally Posted by devonblzx View Post
    These results look strange to me. Personally I have tested with NGINX on my single processor quad-core and have gotten results equal to or greater than what is shown and for a more intense HTML file (>1KB) and much better results for A hello world PHP.

    I think there needs to be more organization with your testing as results are sporadic as of now.
    Everyone is doing this in their spare time for fun. If you think you could do better you're welcome to jump on and run a few tests yourself.
    Ask me about CloudCentrum (coming soon) -- The complete, turn-key cloud software solution

  16. #91
    Join Date
    May 2001
    Location
    Dayton, Ohio
    Posts
    4,977
    Quote Originally Posted by e-Sensibility View Post
    Yes, what about siege instead of ab? Let's give that a try.
    Personally I would do the tests with a few web common web apps with actual data and then use both ab and siege to generate load. I've been using siege lately to load test and I quite like it but ab will always be in my toolkit.

    The nice thing about siege is that you can pass it a file of several URLs and have it try to simulate actual user patterns. This is leads to better real world results. Hitting the front page of a site repeatedly is great for testing attack resistance but very poor for user experience testing.

  17. #92
    Join Date
    Nov 2005
    Posts
    3,944
    Quote Originally Posted by e-Sensibility View Post
    Everyone is doing this in their spare time for fun. If you think you could do better you're welcome to jump on and run a few tests yourself.
    Are you using a common PHP build? I think it would be best to make sure all tests are using the same php version/modules.

    I'd be happy to jump in and do some testing if you would like. Feel free to email me devon at rockmyweb.net.


  18. #93
    Join Date
    Aug 2004
    Location
    Earth
    Posts
    8,154
    Quote Originally Posted by The Prohacker View Post
    Personally I would do the tests with a few web common web apps with actual data and then use both ab and siege to generate load. I've been using siege lately to load test and I quite like it but ab will always be in my toolkit.

    The nice thing about siege is that you can pass it a file of several URLs and have it try to simulate actual user patterns. This is leads to better real world results. Hitting the front page of a site repeatedly is great for testing attack resistance but very poor for user experience testing.
    That would be a lot more interesting. Real world results are worth more in my opinion.

  19. #94
    Join Date
    Mar 2009
    Posts
    2,222
    Quote Originally Posted by devonblzx View Post
    These results look strange to me. Personally I have tested with NGINX on my single processor quad-core and have gotten results equal to or greater than what is shown and for a more intense HTML file (>1KB) and much better results for A hello world PHP.
    Mmm.

    OK, I've never used nginx; but I downloaded the source, compiled it, and ran it on my 4GB Pentium dual-core E5200 desktop machine running Fedora 11. I test using ab and 127.0.0.1

    With a single ab instance, I got

    5000-6000 requests a second: ab -n 30000 127.0.0.1:8181/test.html
    10000-11000 requests a second: ab -k -n 30000 127.0.0.1:8181/test.html
    14000-20000 requests a second: ab -c 100 -k -n 30000 127.0.0.1:8181/test.html

    Here's the output from the 20,000 requests-a-second run

    $ ab -c 100 -k -n 30000 127.0.0.1:8181/test.html
    This is ApacheBench, Version 2.3 <$Revision: 655654 $>
    Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
    Licensed to The Apache Software Foundation, http://www.apache.org/

    Benchmarking 127.0.0.1 (be patient)
    Completed 3000 requests
    Completed 6000 requests
    Completed 9000 requests
    Completed 12000 requests
    Completed 15000 requests
    Completed 18000 requests
    Completed 21000 requests
    Completed 24000 requests
    Completed 27000 requests
    Completed 30000 requests
    Finished 30000 requests


    Server Software: nginx/0.7.64
    Server Hostname: 127.0.0.1
    Server Port: 8181

    Document Path: /test.html
    Document Length: 101 bytes

    Concurrency Level: 100
    Time taken for tests: 1.467 seconds
    Complete requests: 30000
    Failed requests: 0
    Write errors: 0
    Keep-Alive requests: 29748
    Total transferred: 9509374 bytes
    HTML transferred: 3030202 bytes
    Requests per second: 20450.01 [#/sec] (mean)
    Time per request: 4.890 [ms] (mean)
    Time per request: 0.049 [ms] (mean, across all concurrent requests)
    Transfer rate: 6330.30 [Kbytes/sec] received

    Connection Times (ms)
    min mean[+/-sd] median max
    Connect: 0 0 1.0 0 23
    Processing: 0 5 6.6 4 118
    Waiting: 0 5 6.6 4 118
    Total: 0 5 6.7 4 119

    Percentage of the requests served within a certain time (ms)
    50% 4
    66% 5
    75% 6
    80% 7
    90% 8
    95% 9
    98% 15
    99% 23
    100% 119 (longest request)

  20. #95
    Join Date
    Nov 2005
    Posts
    3,944
    Yes, I'm guessing there is either something configured wrong with the previous nginx test or some type of bottleneck. It is also the developmental version he used, not the stable. I'll hop on the server and run the test once given access.

  21. #96
    Quote Originally Posted by devonblzx View Post
    Yes, I'm guessing there is either something configured wrong with the previous nginx test or some type of bottleneck. It is also the developmental version he used, not the stable. I'll hop on the server and run the test once given access.
    PM e-Sensibility to get your access.
    LiteSpeed Web Acceleration Platform by https://www.litespeedtech.com
    Apache drop-in replacement. Triple server capacity with 10X performance increases.
    Ultimate web serving platform for WordPress, Magento and other web applications.
    Turbo charging all WordPress sites hosted on your server with a single click!

  22. #97
    5000-6000 requests a second: ab -n 30000 127.0.0.1:8181/test.html
    10000-11000 requests a second: ab -k -n 30000 127.0.0.1:8181/test.html
    14000-20000 requests a second: ab -c 100 -k -n 30000 127.0.0.1:8181/test.html
    Do not mix keepalive and non-keepalive results together, we only compared the non-keepalive results from lighttpd, nginx and LiteSpeed, all are in the neighborhood 7000-8000 for single ab client test.

    For keepalive test on the test server, nginx does around 25000 reqs/sec, lighttpd does around 9000 reqs/sec, LiteSpeed does around 30000 reqs/sec for single ab client test.
    LiteSpeed Web Acceleration Platform by https://www.litespeedtech.com
    Apache drop-in replacement. Triple server capacity with 10X performance increases.
    Ultimate web serving platform for WordPress, Magento and other web applications.
    Turbo charging all WordPress sites hosted on your server with a single click!

  23. #98
    We have improved LiteSpeed non-keepalive performance by another 10%, so I expect LiteSpeed can deliver around 8500 req/s on the test server with one ab client.

    However, the test server is not accessible now, so I cannot confirm the number.

    e-Sensibility,

    Are we going to continue finishing the showdown?
    LiteSpeed Web Acceleration Platform by https://www.litespeedtech.com
    Apache drop-in replacement. Triple server capacity with 10X performance increases.
    Ultimate web serving platform for WordPress, Magento and other web applications.
    Turbo charging all WordPress sites hosted on your server with a single click!

  24. #99
    Join Date
    Jan 2005
    Posts
    2,203
    Quote Originally Posted by mistwang View Post
    We have improved LiteSpeed non-keepalive performance by another 10%, so I expect LiteSpeed can deliver around 8500 req/s on the test server with one ab client.

    However, the test server is not accessible now, so I cannot confirm the number.

    e-Sensibility,

    Are we going to continue finishing the showdown?
    Is that implemented in the latest 4.0 or 4.1 release?

  25. #100
    Join Date
    Jan 2005
    Posts
    2,203
    So, no one wants to update this?

Page 4 of 6 FirstFirst 123456 LastLast

Similar Threads

  1. Apache vs Litespeed
    By Alan108 in forum Web Hosting
    Replies: 4
    Last Post: 10-08-2009, 09:54 PM
  2. LiteSpeed Vs Apache?
    By leanfarrell in forum Hosting Security and Technology
    Replies: 11
    Last Post: 10-06-2009, 03:25 PM
  3. LiteSpeed vs Apache
    By LadySDevil in forum Hosting Security and Technology
    Replies: 51
    Last Post: 07-21-2009, 12:24 AM
  4. LiteSpeed -> Apache
    By goooh in forum Hosting Security and Technology
    Replies: 6
    Last Post: 07-13-2009, 12:26 AM
  5. apache vs litespeed
    By linktome in forum Hosting Software and Control Panels
    Replies: 1
    Last Post: 07-13-2008, 08:37 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •