Here’s the code:

import time, math

t = time.asctime() print t x = 1 for i in range(1000000): x = math.tanh(i) t = time.asctime() print t

With the DH installed version on the server dali Python 2.1.3 (#1, Sep 7 2002, 15:29:56) [GCC 2.95.4 20011002 (Debian prerelease)] on linux2 I got a time of five seconds on five straight runs.

With my build of python 2.4 on the same server: Python 2.4 (#1, Dec 14 2004, 00:04:38) [GCC 2.95.4 20011002 (Debian prerelease)] on linux2 I got times of three seconds on every run but one, which took six seconds. I alternated between versions so as to hopefully mitigate any effects of the shared server being more or less busy at different times.

Despite the fact that the mathematical performance does seem to be improved, I have also noticed that python 2.4 seems to be slower to start up.

Unfortunately, that obviously causes a big hit on cgi performance for us, since we don’t have access to mod_python on a DH shared server.

The built-in python is almost instantaneous to display the interpreter prompt, but my version of python 2.4 can take anywhere from almost instantaneous (maybe a 1/4 second) up to a second.

For example, just entering the python interactive mode takes several seconds in 2.4. Any ideas on how to speed this up?