Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.python > #2578

Re: Python CPU

Show all headers | View raw

On 4/4/2011 5:23 AM, Paul Rubin wrote:
> Gregory Ewing<greg.ewing@canterbury.ac.nz>  writes:
>> What might help more is having bytecodes that operate on
>> arrays of unboxed types -- numpy acceleration in hardware.
>
> That is an interesting idea as an array or functools module patch.
> Basically a way to map or fold arbitrary functions over arrays, with a
> few obvious optimizations to avoid refcount churning.  It could have
> helped with a number of things I've done over the years.

For map, I presume you are thinking of an array.map(func) in system code 
(C for CPython) equivalent to

def map(self,func):
   for i,ob in enumerate(self):
     self[i] = func(ob)

The question is whether it would be enough faster. Of course, what would 
really be needed for speed are wrapped system-coded funcs that map would 
recognize and pass and received unboxed array units to and from. At that 
point, we just about invented 1-D numpy ;-).

I have always thought the array was underutilized, but I see now that it 
only offers Python code space saving at a cost of interconversion time. 
To be really useful, arrays of unboxed data, like strings and bytes, 
need system-coded functions that directly operate on the unboxed data, 
like strings and bytes have. Array comes with a few, but very few, 
generic sequence methods, like .count(x) (a special-case of reduction).

-- 
Terry Jan Reedy

Back to comp.lang.python | Previous | Next — Previous in thread | Next in thread | Find similar

Thread

Python CPU Brad <hwfwguy@gmail.com> - 2011-04-01 08:38 -0700
  Re: Python CPU jkn <jkn_gg@nicorp.f9.co.uk> - 2011-04-01 08:59 -0700
  Re: Python CPU Nobody <nobody@nowhere.com> - 2011-04-01 17:52 +0100
    Re: Python CPU Stefan Behnel <stefan_ml@behnel.de> - 2011-04-01 19:00 +0200
    Re: Python CPU geremy condra <debatem1@gmail.com> - 2011-04-01 10:12 -0700
  Re: Python CPU Emile van Sebille <emile@fenx.com> - 2011-04-01 11:28 -0700
  Re: Python CPU Emile van Sebille <emile@fenx.com> - 2011-04-01 11:35 -0700
    Re: Python CPU John Nagle <nagle@animats.com> - 2011-04-01 12:18 -0700
  Re: Python CPU "BartC" <bc@freeuk.com> - 2011-04-02 23:06 +0100
    Re: Python CPU Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2011-04-03 04:03 +0000
  Re: Python CPU Gregory Ewing <greg.ewing@canterbury.ac.nz> - 2011-04-03 12:10 +1200
    Re: Python CPU Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2011-04-03 04:01 +0000
      Re: Python CPU John Nagle <nagle@animats.com> - 2011-04-03 00:16 -0700
        Re: Python CPU Paul Rubin <no.email@nospam.invalid> - 2011-04-03 01:59 -0700
          Re: Python CPU Gregory Ewing <greg.ewing@canterbury.ac.nz> - 2011-04-04 19:41 +1200
        Re: Python CPU Gregory Ewing <greg.ewing@canterbury.ac.nz> - 2011-04-04 19:47 +1200
          Re: Python CPU Paul Rubin <no.email@nospam.invalid> - 2011-04-04 02:23 -0700
            Re: Python CPU Terry Reedy <tjreedy@udel.edu> - 2011-04-04 13:14 -0400
            Re: Python CPU Terry Reedy <tjreedy@udel.edu> - 2011-04-04 14:58 -0400
              Re: Python CPU Gregory Ewing <greg.ewing@canterbury.ac.nz> - 2011-04-05 09:59 +1200
          Re: Python CPU geremy condra <debatem1@gmail.com> - 2011-04-04 09:36 -0700
            Re: Python CPU Gregory Ewing <greg.ewing@canterbury.ac.nz> - 2011-04-05 10:05 +1200
          Re: Python CPU John Nagle <nagle@animats.com> - 2011-04-04 12:56 -0700
            Re: Python CPU Paul Rubin <no.email@nospam.invalid> - 2011-04-04 13:20 -0700
              Re: Python CPU Gregory Ewing <greg.ewing@canterbury.ac.nz> - 2011-04-05 10:11 +1200
    Re: Python CPU Werner Thie <werner@thieprojects.ch> - 2011-04-03 17:44 +0200
      Re: Python CPU John Nagle <nagle@animats.com> - 2011-04-03 10:15 -0700
        Re: Python CPU Nobody <nobody@nowhere.com> - 2011-04-03 22:37 +0100
          Re: Python CPU Dennis Lee Bieber <wlfraed@ix.netcom.com> - 2011-04-03 22:59 -0700

csiph-web