From: Valery Pipin <pip@iszf.irk.ru>
To: ALT Linux Team development discussions <devel@lists.altlinux.org>
Subject: Re: [devel] atlas is broken again?
Date: Tue, 5 Apr 2011 01:02:17 -0700
Message-ID: <201104050102.18009.pip@iszf.irk.ru> (raw)
In-Reply-To: <4D9AC7A7.1030306@mmedia2.kemsu.ru>
В сообщении от 5 апреля 2011 00:41:27 автор REAL написал:
> 05.04.2011 14:34, Valery Pipin пишет:
> >> Из сизифа?
> >
> > Отсюда, http://www.enthought.com/
>
> А зачем, если есть в сизифе?
>
сизифовский медленный, и к тому же периодически ломается. В enthought сборке
использована MKL и линейная алгебра реально паралеллится если есть куда.
> >> покажите rpm -qi libatlas
> >
> > я писал что я вернулся е прежнюю, рабочую, версию
>
> Тогда я вообще в растерянности, ибо ни провала в производительности,
> ни в падении ничего не заметил.
Возвращаемся к первому сообщению. Вот это с libatlas-3.9.35:
IPython 0.10 -- An enhanced Interactive Python.
? -> Introduction and overview of IPython's features.
%quickref -> Quick reference.
help -> Python's own help system.
object? -> Details about 'object'. ?object also works, ?? prints more.
In [1]: from numpy.random import *
In [2]: from numpy.linalg import *
In [3]: m=randn(2000,2000)
In [4]: %time U,s,V=svd(m)
Ошибка сегментирования
Теперь пробуем для массивов меньших размеров:
In [1]: from numpy.linalg import *
In [2]: from numpy.random import *
In [3]: m=randn(100,100)
In [4]: %time U,s,V=svd(m)
CPU times: user 0.01 s, sys: 0.00 s, total: 0.01 s
Wall time: 0.01 s
In [6]: m=randn(200,200)
In [7]: %time U,s,V=svd(m)
---------------------------------------------------------------------------
LinAlgError Traceback (most recent call last)
/home/vv/<ipython console> in <module>()
/usr/lib/python2.6/site-packages/IPython/iplib.pyc in ipmagic(self, arg_s)
1180 else:
1181 magic_args = self.var_expand(magic_args,1)
-> 1182 return fn(magic_args)
1183
1184 def ipalias(self,arg_s):
/usr/lib/python2.6/site-packages/IPython/Magic.pyc in magic_time(self,
parameter_s)
1969 else:
1970 st = clk()
-> 1971 exec code in glob
1972 end = clk()
1973 out = None
/home/vv/<timed exec> in <module>()
/usr/lib64/python2.6/site-packages/numpy/linalg/linalg.pyc in svd(a,
full_matrices, compute_uv)
1318 work, lwork, iwork, 0)
1319 if results['info'] > 0:
-> 1320 raise LinAlgError, 'SVD did not converge'
1321 s = s.astype(_realType(result_t))
1322 if compute_uv:
LinAlgError: SVD did not converge
next prev parent reply other threads:[~2011-04-05 8:02 UTC|newest]
Thread overview: 59+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-04-05 3:44 Valery Pipin
2011-04-05 3:59 ` REAL
2011-04-05 4:20 ` Valery Pipin
2011-04-05 6:19 ` REAL
2011-04-05 6:41 ` Valery Pipin
2011-04-05 6:53 ` REAL
2011-04-05 7:02 ` Valery Pipin
2011-04-05 7:14 ` REAL
2011-04-05 7:34 ` Valery Pipin
2011-04-05 7:41 ` REAL
2011-04-05 8:02 ` Valery Pipin [this message]
2011-04-05 8:20 ` REAL
2011-04-05 17:51 ` Valery Pipin
2011-04-06 3:22 ` REAL
2011-04-06 4:54 ` Valery Pipin
2011-04-06 5:02 ` REAL
2011-04-06 5:04 ` Valery Pipin
2011-04-06 5:15 ` REAL
2011-04-06 5:49 ` Valery Pipin
2011-04-06 6:05 ` REAL
2011-04-06 7:28 ` REAL
2011-04-06 8:36 ` Valery Pipin
2011-04-06 8:47 ` REAL
2011-04-06 8:53 ` REAL
2011-04-06 15:47 ` Valery Pipin
2011-04-06 16:05 ` REAL
2011-04-08 1:52 ` Valery Pipin
2011-04-08 4:07 ` REAL
2011-04-08 4:17 ` Valery Pipin
2011-04-08 5:09 ` REAL
2011-04-08 5:31 ` Valery Pipin
2011-04-08 6:09 ` REAL
2011-04-08 6:44 ` Valery Pipin
2011-04-08 7:08 ` REAL
2011-04-08 8:25 ` REAL
2011-04-09 1:09 ` Valery Pipin
2011-04-09 1:19 ` REAL
2011-04-09 1:52 ` Valery Pipin
2011-04-09 2:08 ` REAL
2011-04-09 9:32 ` Michael Shigorin
2011-04-09 11:22 ` REAL
2011-04-10 1:42 ` Valery Pipin
2011-04-08 5:11 ` REAL
2011-04-08 4:24 ` REAL
2011-04-08 5:43 ` Anton Farygin
2011-04-08 5:47 ` REAL
2011-04-08 6:11 ` Anton Farygin
2011-04-08 6:20 ` REAL
2011-04-08 10:01 ` REAL
2011-04-08 10:05 ` REAL
2011-04-08 10:08 ` Anton Farygin
2011-04-08 10:35 ` REAL
2011-04-08 10:53 ` REAL
2011-04-08 11:00 ` REAL
2011-04-08 11:01 ` Anton Farygin
2011-04-08 11:12 ` REAL
2011-04-06 8:19 ` REAL
2011-04-07 4:54 ` REAL
2011-04-07 5:20 ` Valery Pipin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=201104050102.18009.pip@iszf.irk.ru \
--to=pip@iszf.irk.ru \
--cc=devel@lists.altlinux.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
ALT Linux Team development discussions
This inbox may be cloned and mirrored by anyone:
git clone --mirror http://lore.altlinux.org/devel/0 devel/git/0.git
# If you have public-inbox 1.1+ installed, you may
# initialize and index your mirror using the following commands:
public-inbox-init -V2 devel devel/ http://lore.altlinux.org/devel \
devel@altlinux.org devel@altlinux.ru devel@lists.altlinux.org devel@lists.altlinux.ru devel@linux.iplabs.ru mandrake-russian@linuxteam.iplabs.ru sisyphus@linuxteam.iplabs.ru
public-inbox-index devel
Example config snippet for mirrors.
Newsgroup available over NNTP:
nntp://lore.altlinux.org/org.altlinux.lists.devel
AGPL code for this site: git clone https://public-inbox.org/public-inbox.git