Timing results

modified one

[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/one.py                      (05-31 16:46)
./python /tmp/one.py  10.62s user 0.02s system 97% cpu 10.955 total
[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/one.py                      (05-31 16:47)
./python /tmp/one.py  10.32s user 0.01s system 97% cpu 10.627 total
[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/one.py                      (05-31 16:47)
./python /tmp/one.py  10.47s user 0.01s system 97% cpu 10.781 total
[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/one.py                      (05-31 16:47)
./python /tmp/one.py  10.72s user 0.01s system 95% cpu 11.212 total
[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/one.py                      (05-31 16:47)
./python /tmp/one.py  10.82s user 0.01s system 97% cpu 11.138 total

modified two

[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/two.py                      (05-31 16:52)
./python /tmp/two.py  24.60s user 13.31s system 97% cpu 39.050 total
[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/two.py                      (05-31 16:53)
./python /tmp/two.py  25.67s user 13.35s system 92% cpu 42.319 total
[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/two.py                      (05-31 16:54)
./python /tmp/two.py  25.49s user 13.13s system 97% cpu 39.742 total
[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/two.py                      (05-31 16:54)
./python /tmp/two.py  26.51s user 13.50s system 97% cpu 41.181 total
[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/two.py                      (05-31 16:55)
./python /tmp/two.py  26.14s user 14.25s system 96% cpu 41.738 total

unmodified one

[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/one.py                      (05-31 17:03)
./python /tmp/one.py  10.66s user 0.02s system 95% cpu 11.182 total
[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/one.py                      (05-31 17:04)
./python /tmp/one.py  10.89s user 0.03s system 95% cpu 11.422 total
[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/one.py                      (05-31 17:04)
./python /tmp/one.py  10.55s user 0.02s system 97% cpu 10.890 total
[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/one.py                      (05-31 17:04)
./python /tmp/one.py  10.95s user 0.01s system 97% cpu 11.293 total
[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/one.py                      (05-31 17:05)
./python /tmp/one.py  10.55s user 0.02s system 96% cpu 10.944 total

unmodified two

[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/two.py                      (05-31 17:07)
./python /tmp/two.py  27.91s user 14.11s system 96% cpu 43.399 total
[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/two.py                      (05-31 17:08)
./python /tmp/two.py  26.74s user 13.62s system 97% cpu 41.515 total
[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/two.py                      (05-31 17:09)
./python /tmp/two.py  27.91s user 13.80s system 96% cpu 43.195 total
[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/two.py                      (05-31 17:09)
./python /tmp/two.py  25.19s user 13.83s system 97% cpu 40.157 total
[kb1pkl@Ulysses:src/Python-3.2.3]$ time ./python /tmp/two.py                      (05-31 17:10)
./python /tmp/two.py  27.56s user 14.08s system 96% cpu 43.049 total

the funny thing is that no one expects gethostbyname to be highly efficient. it gets called one in the lifetime of every socket -- negligible compared to send/recv