public inbox for gentoo-science@lists.gentoo.org
 help / color / mirror / Atom feed
From: "M. Edward (Ed) Borasky" <znmeb@cesmail.net>
To: gentoo-science@lists.gentoo.org
Subject: Re: [gentoo-science] lapack 3.0? (rkward)
Date: Tue, 27 Feb 2007 07:17:24 -0800	[thread overview]
Message-ID: <45E44B84.9070401@cesmail.net> (raw)
In-Reply-To: <45E3E7E3.1090507@centrum.cz>

Honza Macháček wrote:
> On 26.02.2007 01:38, M. Edward (Ed) Borasky wrote:
>   
>> I'm a little curious about something ... does "rkward" actually require
>> an external Lapack, or is this optional? The reason I ask is that R
>> itself *contains* a copy of Lapack, and the R developers have deprecated
>> linking against external Lapack libraries!
>>     
>
>   I think I recall some discussion at the bugzilla to the point that all
> multimedia packages that compile their own libraries instead of linking
> to external ones should be hardmasked, allegedly for the sake of
> maintainability. Moreover in my mailbox I've found an old discussion
> (28. 8. 2006) of Freemat 2.0 and problems with its dependency on matio,
> when accidentally I myself called to attention that Freemat sources
> contained their own copy of the matio library, and Andrey G. Grozin
> dismissed the option of using them rather fiercely, reasoning:
>
>   
>> This goes against modularity, *the* fundamental principle of any good 
>> Linux distribution.
>>     
Well, here's the logic of the R developers (from 
http://cran.fhcrc.org/doc/manuals/R-admin.html):


        "A.3.2 LAPACK

Provision is made for using an external LAPACK library, principally to 
cope with BLAS libraries which contain a copy of LAPACK (such as 
|libsunperf| on Solaris, |vecLib| on Mac OS X and ACML on 
`ix86'/`x86_64' Linux). However, the likely performance gains are 
thought to be small (and may be negative), and the default is not to 
search for a suitable LAPACK library, and this is definitely *not* 
recommended.

[snip]

"For problems compiling LAPACK using recent versions of |gcc| on `ix86' 
Linux, see New platforms 
<http://cran.fhcrc.org/doc/manuals/R-admin.html#New-platforms>: these 
problems have surfaced in Fedora Core 3's distribution, for example.

"Please *do* bear in mind that using --with-lapack is `definitely *not* 
recommended': it is provided *only* because it is necessary on some 
platforms."

In any event, when I have beta-tested R releases on my Athlon Thunderbird, when I use the "with-lapack" option and link R with the external Atlas LAPACK I get many failures in the "make check-all" pass. These don't happen when I just use the Atlas BLAS, so I don't use the external LAPACK with R.
-- 
M. Edward (Ed) Borasky, FBG, AB, PTA, PGS, MS, MNLP, NST, ACMC(P)
http://borasky-research.blogspot.com/

If God had meant for carrots to be eaten cooked, He would have given rabbits fire.

-- 
gentoo-science@gentoo.org mailing list



      reply	other threads:[~2007-02-27 15:18 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-02-24 13:07 [gentoo-science] lapack 3.0? Matthew R. Lee
2007-02-24 21:50 ` Patrice Beaud
2007-02-25 20:18   ` Matthew R. Lee
2007-02-26  0:38     ` M. Edward (Ed) Borasky
2007-02-26 13:09       ` Matthew R. Lee
2007-02-27  8:12       ` [gentoo-science] lapack 3.0? (rkward) Honza Macháček
2007-02-27 15:17         ` M. Edward (Ed) Borasky [this message]

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=45E44B84.9070401@cesmail.net \
    --to=znmeb@cesmail.net \
    --cc=gentoo-science@lists.gentoo.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
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox