CODEWITHSUNDEEP
phpapachebenchmarking
Marty Wallace
Marty Wallace

Reputation: 35734

Apache benchmark inconsistent

I am testing my website homepage on a local server using apache benchmark and I would have expected some consistent results, given that there are no eternal influences on the site as it is local.

However, I am getting wildly different results, can anyone explain this:

ab -n 10 -c 1 -g benchmark-data.txt http://dev.local/

-

starttime       seconds ctime   dtime   ttime   wait
Mon Nov 26 13:53:49 2012        1353938029      0       693     693     670
Mon Nov 26 13:53:52 2012        1353938032      0       891     891     865
Mon Nov 26 13:53:48 2012        1353938028      0       975     975     950
Mon Nov 26 13:53:51 2012        1353938031      0       1212    1212    1191
Mon Nov 26 13:53:53 2012        1353938033      0       1344    1344    1324
Mon Nov 26 13:53:45 2012        1353938025      0       1352    1353    1335
Mon Nov 26 13:53:54 2012        1353938034      0       1398    1398    1379
Mon Nov 26 13:53:49 2012        1353938029      0       1409    1409    1391
Mon Nov 26 13:53:55 2012        1353938035      0       1430    1430    1411
Mon Nov 26 13:53:46 2012        1353938026      0       1509    1509    1489



starttime       seconds ctime   dtime   ttime   wait
Mon Nov 26 14:01:51 2012        1353938511      0       515     515     497
Mon Nov 26 14:01:49 2012        1353938509      0       685     685     667
Mon Nov 26 14:01:49 2012        1353938509      0       752     752     731
Mon Nov 26 14:01:51 2012        1353938511      0       870     870     848
Mon Nov 26 14:01:50 2012        1353938510      0       872     872     855
Mon Nov 26 14:01:48 2012        1353938508      0       894     894     873
Mon Nov 26 14:01:47 2012        1353938507      0       902     902     883
Mon Nov 26 14:01:45 2012        1353938505      0       933     933     909
Mon Nov 26 14:01:46 2012        1353938506      0       1202    1202    1184
Mon Nov 26 14:01:42 2012        1353938502      0       3009    3009    2991


starttime       seconds ctime   dtime   ttime   wait
Mon Nov 26 14:02:45 2012        1353938565      0       916     917     892
Mon Nov 26 14:02:40 2012        1353938560      0       994     994     974
Mon Nov 26 14:02:44 2012        1353938564      0       1032    1032    1009
Mon Nov 26 14:02:39 2012        1353938559      0       1255    1255    1235
Mon Nov 26 14:02:37 2012        1353938557      0       1319    1319    1302
Mon Nov 26 14:02:36 2012        1353938556      0       1355    1355    1338
Mon Nov 26 14:02:41 2012        1353938561      0       1375    1376    1350
Mon Nov 26 14:02:48 2012        1353938568      0       1429    1429    1405
Mon Nov 26 14:02:46 2012        1353938566      0       1689    1689    1663
Mon Nov 26 14:02:42 2012        1353938562      0       1982    1982    1964

Upvotes: 1

Views: 224

Answers (1)

Sudar
Sudar

Reputation: 19982

Even though the chance of external interference is low, because you are running the site in localhost, it is still possible to get varying results because of other factors like

  • The current load on your machine
  • Memory usage by other apps

etc.

Also try to run the benchmark with higher values of n. Something like -n100 or may be even -n1000

Upvotes: 1

Related Questions