[cpu minus startup time]     [sort] [sort]   Source Code CPU (sec) Mem (KB) Lines Code Log gcc 0.0230 364 17 log out se 0.0240 412 45 log out g++ 0.0240 828 14 log out bigloo 0.0390 940 18 log out stalin 0.0390 720 23 log out ocaml 0.0430 540 6 log out oberon2 0.0480 1160 19 log out clean 0.0530 604 18 log out mlton 0.0660 628 25 log out slang 0.0780 880 9 log out felix 0.0830 872 19 log out gnat 0.1700 1080 29 log out cmucl 0.1740 4632 17 log out sbcl 0.1940 5672 17 log out nice 0.2010 5388 9 log out gcj 0.2020 5504 13 log out chicken 0.2200 1036 18 log out smlnj 0.2440 1256 36 log out java 0.2700 7212 13 log out kaffe 0.3080 6060 13 log out poly 0.4619 2416 39 log out gforth 0.5009 808 11 log out ghc 1.4548 1036 21 log out sablevm 1.8157 2900 13 log out pop11 2.3626 4952 23 log out fpascal 2.5196 296 0 log out ocamlb 2.5646 752 6 log out gij 2.8216 10400 13 log out pike 3.2265 2984 11 log out oz 3.2915 2640 25 log out rep 3.4865 1004 16 log out mercury 3.8714 5196 38 log out gst 4.0584 3904 7 log out lua 4.4093 1376 13 log out xemacs 5.4442 7760 13 log out hipe 5.5722 4440 26 log out python 6.1531 2300 12 log out psyco 6.1951 2300 14 log out erlang 6.2011 4412 26 log out poplisp 7.8378 13896 15 log out mzc 9.0546 3380 18 log out icon 9.4706 2000 11 log out perl 9.5395 1900 10 log out newlisp 10.2334 1204 15 log out guile 17.3774 1964 15 log out php 19.7570 4744 9 log out mzscheme 20.3479 3400 18 log out tcl 20.3649 3456 12 log out gawk 25.4801 2128 8 log out ruby 26.8429 1556 12 log out mawk 27.1859 1180 8 log out nhc98 288.7291 156804 13 log out Languages that compile to native code are in Bold Italics.

[Note: Values have been normalized to fall in the range of 0-10 for aesthetic reasons. Original value ranges are included on the X-axis.] Click here for more detailed data and graphs. [Results last updated: Fri Oct 15 23:56:42 2004 CDT]

For this test, each program should be implemented in the same way.

Originally this test was pretty much the same as the array test from Timing Trials, or, the Trials of Timing: Experiments with Scripting and User-Interface Languages by Brian W. Kernighan and Christopher J. Van Wyk. My original implementation of this test was pretty close to the Kernighan and Wyk version. I also did an experiment of manually unrolling the loop around the array. These two versions of the test are now obsoleted and superceded by this page.

This test is supposed to just test array access via subscripting. However, the original test didn't take into account array initilization overhead, and so in this version we create an array and then re-use it many times.

Note that one loop counts down from the end of the array to the beginning. This is to give the advantage to a random-access data structure, as an array should be, and not just any sequence data structure, as such as a list.

  1000 1000000