Re: python3.0 - any hope it will get faster?
On 9 Dez., 11:51, Helmut Jarausch [EMAIL PROTECTED] wrote: Hi, I was somewhat surprised when I ran pystones with python-2.5.2 and with python-3.0 On my old/slow machine I get python-2.5.2 from test import pystone pystone.pystones() gives (2.73, 18315.018315018315) python-3.0 from test import pystone pystone.pystones() gives (4.2705, 11709.601873536298) That's a drop of 36% ! I know that processing unicode is inherently slower, but still I was surprised that it's so much slower. On my WinXP notebook python-3.0 from test import pystone pystone.pystones() (1.1734318188484849, 42610.059823557647) python-2.5.1 from test import pystone pystone.pystones() (1.2927221197107452, 38678.072601703308) -- http://mail.python.org/mailman/listinfo/python-list
Re: python3.0 - any hope it will get faster?
Helmut Jarausch wrote: I know that processing unicode is inherently slower, but still I was surprised that it's so much slower. Is there any hope Python-3.0 will get faster or is the main potential for optimizations exhausted, already? That's not to start a flame war! I know computers get faster, we human beings don't (me, at least) Don't worry, it's going to get faster. Our top priority was feature completeness and stability. With the first final version out we are focusing on bug fixes and speed ups. We have a fair bunch of speed patches and ideas in our work queue. For example Mark and Victor are working on long integer optimizations, Alexandre and I are focusing on the new IO library and so on. Some of the speedup will make it into 3.0.1 and future patch releases of the 3.0 series. Larger and more complex chances can be expected for Python 3.1. There are interesting experiments with LLVM and threaded code (not to confuse with multithreading) going on. Grüße an die andere Seite von Aachen Christian -- http://mail.python.org/mailman/listinfo/python-list