Path: csiph.com!usenet.pasdenom.info!news.albasani.net!feeder.erje.net!newsfeed.xs4all.nl!newsfeed6.news.xs4all.nl!xs4all!post.news.xs4all.nl!not-for-mail
Newsgroups: comp.lang.python
Date: Wed, 17 Oct 2012 07:45:50 -0700 (PDT)
In-Reply-To: <mailman.2348.1350484621.27098.python-list@python.org>
Complaints-To: groups-abuse@google.com
Injection-Info: glegroupsg2000goo.googlegroups.com; posting-host=123.192.32.215; posting-account=5JdMBQoAAABHnS4mjpqEzxnmWtgiiVNw
References: <CA+C4C6f=24oD4seHY_-h1F0GMvBTWNYbzTABXOdtKkAZtE9hdQ@mail.gmail.com> <507E145E.9060008@davea.name> <CA+C4C6cjha4bMZ9gyx5fTp0c_nfFQxRGu31dGNMjgHQBu73=ig@mail.gmail.com> <mailman.2341.1350477185.27098.python-list@python.org> <61d76114-3021-4a3b-9943-1e81147d3ce1@6g2000pbh.googlegroups.com> <mailman.2348.1350484621.27098.python-list@python.org>
User-Agent: G2/1.0
MIME-Version: 1.0
Subject: Re: list comprehension question
From: 88888 Dihedral <dihedral88888@googlemail.com>
To: comp.lang.python@googlegroups.com
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Cc: rusi <rustompmody@gmail.com>, d@davea.name, python-list@python.org
Precedence: list
Message-ID: <mailman.2349.1350485158.27098.python-list@python.org>
Lines: 77
NNTP-Posting-Host: 2001:888:2000:d::a6
Xref: csiph.com comp.lang.python:31514

Dave Angel=E6=96=BC 2012=E5=B9=B410=E6=9C=8817=E6=97=A5=E6=98=9F=E6=9C=9F=
=E4=B8=89UTC+8=E4=B8=8B=E5=8D=8810=E6=99=8237=E5=88=8601=E7=A7=92=E5=AF=AB=
=E9=81=93=EF=BC=9A
> On 10/17/2012 10:06 AM, rusi wrote:
>=20
> > On Oct 17, 5:33 pm, Dave Angel <d...@davea.name> wrote:
>=20
> >> On 10/17/2012 12:43 AM, Kevin Anthony wrote:> Is it not true that list=
 comprehension is much faster the the for loops?
>=20
> >>
>=20
> >>> If it is not the correct way of doing this, i appoligize.
>=20
> >>> Like i said, I'm learing list comprehension.
>=20
> >> list comprehensions CAN be much faster, but not necessarily.  The most
>=20
> >> complex a loop, the less likely it'll help much.
>=20
> > One-lining the comprehension seems to make a difference of about 10%
>=20
> > out here. Maybe Ive missed something? Seems too large=C3=AF=C2=BF=C2=BD
>=20
> >
>=20
> > # My original suggestion
>=20
> > def dot(p,q): return sum (x*y for x,y in zip(p,q))
>=20
> > def transpose(m): return zip(*m)
>=20
> > def mm(a,b): return mmt(a, transpose(b))
>=20
> > def mmt(a,b): return [[dot(ra, rb) for rb in b] for ra in a]
>=20
> >
>=20
> > # One-liner (Thanks Hans for reminding me of sum)
>=20
> >
>=20
> > def mm1(a,b): return [[sum([x*y for x,y in zip(ra,rb)]) for rb in
>=20
> > zip(*b)] for ra in a]
>=20
> >
>=20
> >>>> t1=3DTimer("res=3Dmm1(m,m)", setup=3D"from __main__ import mm1, m")
>=20
> >>>> t1.timeit(1000)
>=20
> > 12.276363849639893
>=20
> >>>> t0=3DTimer("res=3Dmm(m,m)", setup=3D"from __main__ import mm, m")
>=20
> >>>> t0.timeit(1000)
>=20
> > 13.453603029251099
>=20
>=20
>=20
> And I'd wager all the improvement is in the inner loop, the dot() functio=
n.
>=20
>=20
>=20
>=20
>=20
> --=20
>=20
>=20
>=20
> DaveA

Thanks for the tips of matrix operations over some fields or rings=20
other than the real field and the complex field.