Next | Making Programs Faster | 158 |
Perhaps counting modules biased the numbers?
Subroutine time %time cum. lines cum. (%) letter_histogram .76 79.2 79.2 3 3 5.9 BEGIN .10 10.4 89.6 13 16 31.4 handle_message .09 9.4 99.0 7 23 45.1 report .01 1.0 100.0 20 43 84.3 from_histogram .00 0.0 100.0 3 46 90.2 pairify .00 0.0 100.0 5 51 100.0
No, we still have 6% of the code accounting for 80% of the run time
Next | Copyright © 2003 M. J. Dominus |