This is the mail archive of the
`gcc-patches@gcc.gnu.org`
mailing list for the GCC project.

Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|

Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |

Other format: | [Raw text] |

*From*: Tobias Burnus <burnus at net-b dot de>*To*: Richard Guenther <richard dot guenther at gmail dot com>*Cc*: gcc-patches <gcc-patches at gcc dot gnu dot org>, "fortran at gcc dot gnu dot org" <fortran at gcc dot gnu dot org>, Gerald Pfeifer <gerald at pfeifer dot com>*Date*: Mon, 23 Oct 2006 11:05:02 +0200*Subject*: Re: [wwwdocs] Add "-fexternal-blas" to gcc-4.3/changes.html*References*: <453BF79A.1040604@net-b.de> <84fc9c000610230030x24c42c29ha458f9ae55f23655@mail.gmail.com>

Hi, Richard Guenther wrote: > Does it work with BLAS only or with any lapack style library? In this > case at least the documentation should read like "calls to a linear > algebra library > like BLAS or LAPACK for matrix". Of course in that case I also > dislike the option name, but > I don't have a particular better idea either. It works with any library, which provides sgemm, dgemm, cgemm and zgemm and supports the same arguments as Netlib.org's BLAS. It thus works with Atlas, AMD Core Math Library, NAG Fortran Library, Intel Mathe Kernel Library, gotoBLAS, Sun Performance Libraray etc. In fact, the idea is that an optimized library is used. For me BLAS and a library, which provides the routines of BLAS, are in this context the same. And thus I also like the option name "-fexternal-blas" (which was not coined by me, by the way). For me at least "a linear algebra library like BLAS or LAPACK" is much harder to gasp. Is now LAPACK or BLAS needed? Answer: only BLAS. A "linear algebra library like" also is rather vague; what routines are needed? But if you think that many people think of only Netlib's BLAS when one uses the term "BLAS", we indeed need to find a better wording. (Here and maybe [or maybe not] also in the manual, were we currently use: http://gcc.gnu.org/viewcvs/trunk/gcc/fortran/invoke.texi?r1=117948&r2=117947&pathrev=117948 "This option will make gfortran generate calls to BLAS functions for some matrix operations like @code{MATMUL}, instead of using our own algorithms, if the size of the matrices involved is larger than a given limit (see @command{-fblas-matmul-limit}). This may be profitable if an optimized vendor BLAS library is available. The BLAS library will have to be specified at link time.") Tobias

**Follow-Ups**:**Re: [wwwdocs] Add "-fexternal-blas" to gcc-4.3/changes.html***From:*Erik Edelmann

**References**:**[wwwdocs] Add "-fexternal-blas" to gcc-4.3/changes.html***From:*Tobias Burnus

**Re: [wwwdocs] Add "-fexternal-blas" to gcc-4.3/changes.html***From:*Richard Guenther

Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|

Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |