On Tue, Jan 5, 2021 at 3:44 PM Patrik Schindler <poc@xxxxxxxxxx> wrote:
Am 05.01.2021 um 19:44 schrieb John Yeung <gallium.arsenide@xxxxxxxxx>:
iSeriesPython 2.7 runs on it! :)
Seen relatively, Python stays slow, even on fast machines. ;-) (Compared to Perl, for example.)
All benchmarks I have seen show that Python and Perl are both quite
slow on conventional platforms. Perl is faster in some areas, Python
faster in others. They are close enough that it is impossible to say
definitively one is faster than the other for your particular use case
without actually implementing your use case in both languages, and
then test them on your own machine. But overall, Python is usually
found to be slightly *faster* than Perl.
If you are using any of Python, Perl, PHP, or Ruby, you are pretty
much conceding that performance is not the top priority.
Perl is particularly fast for regex processing, which is presumably
what it was originally designed for, mainly. In most other aspects,
it's slower than its competitors.
Some example comparisons:
https://blog.famzah.net/2016/02/09/cpp-vs-python-vs-perl-vs-php-performance-benchmark-2016/
https://stuffivelearned.org/doku.php?id=programming:general:phpvspythonvsperl
Note that iSeriesPython is a special animal, which should be of
particular interest to hobbyists, because it is implemented in ILE C
and runs in the i5/OS environment. Today's Perl and Python for PASE
are "merely" lightly ported Unix versions.
Also note that IBM midrange is not included in what I called
"conventional platforms". I have noticed that even during off-peak
hours, when the machine is "quiet", our IBM i runs Python vastly
slower than the PC on my desk, even though in many respects, the IBM i
is clearly vastly more powerful than my PC.
John Y.
midrange.com
