[Python-Dev] Re: Who cares about the performance of these opcodes?

[Bob Ippolito]

On Mar 10, 2004, at 12:10 PM, Michael Hudson wrote:

> "A.M. Kuchling" <amk at amk.ca> writes:
>
>> On Tue, 09 Mar 2004 08:59:52 -0500, Phillip J. Eby
>> <pje at telecommunity.com> wrote:
>>> I personally don't think it'll help much, if the goal is to reduce
>>> cache misses.  After all, the code is all still there.  But, it
>>> should not do
>>
>> For a planned PyCon lightning talk, I'm benchmarking various
>> combinations of optimizer options.
>> One interesting result: CVS Python gets 25997 pystones on my machine
>> when compiled with
>> -O3 (the default), but 26707 when compiled with gcc's -Os flag.  -Os
>>  optimizes for size,
>> running the subset of the -O2 optimizations that don't increase code 
>> size.
>
> What architecture?  I played around on my ibook with various
> compilation options and running with -fprofile-arcs and so on and
> basically came to the conclusion that nothing made very much
> difference (once past -O2).  Can't remember if I tried -Os.

If you really want faster code you should tell the compiler about the 
particular architecture you need to run it on.  For example, Apple's 
gcc 3.3 has an optimization flag named "-fast" that will (supposedly) 
produce fast non-PIC code that will only run on 64bit G5 processors 
(but can be scaled back to G4 with -mcpu=7450, and probably back to G3 
in a similar way).

-bob