For Windows, we are providing integration with Visual Studio, nmake
with Fortran or PGI compilers. There is no precompiled libs with g77.
We did not look into g77 with MinGW, but I guess if you take the
native Linux package that should work without too much problems if you
stay with Fortran all the way with LAPACK. (I know there is no problem
I guess you have a code in C, no? Usually this is when the mangling
Maybe what you can do is use the CLAPACK package and go with C all the
Also if you are stuck, you can try to post on the LAPACK forum, you
may find a user in the same situation. (http://icl.cs.utk.edu/lapack-forum/
Hope it helps
On Jul 23, 2009, at 6:24 AM, Wilm Schumacher wrote:
ich have a suggestion and a question regarding "LAPACK for Windows".
I have a
really hard time to link the precompiled libs. I think it's a name
problem. I use WindowsXP and g77.exe from MinGW. It is not possible
BLAS.lib because there is the "undefinded reference to <every possible
subroutine>". I'm very new to the topic of name mangling but I think
problem. As I did a "nm" on the libraries I saw that the subroutines
decorated UPPER case and with a LEADING underscore and no underscore
end. This was very ... confusing to me because I read in the
"The GNU g77 compiler converts identifiers to lower case plus an
("foo_"), except that identifiers already containing an underscore
has two underscores appended ("foo_bar__"), following a convention
by f2c. Many other compilers, including SGI's IRIX compilers,
Intel's Fortran compiler, convert all identifiers to lower case plus
underscore ("foo_" and "foo_bar_")"
but found in the output of nm:
00000000 i .drectve
00000000 t .text
00000001 a @feat.00
00000000 T _DAXPY
as an example. As I can see, your name mangling is in another way.
Did you use
the PGI compiler? Did you change the name mangling? If so: Why and
It would by nice and necessary that you explain your name mangling
Furthermore it would be very nice if you explain how to use your
Greetings and Thanks for the good work
ps: If you did the suggested work and I just didn't find it I'm
sorry and would
be very happy about a link.
This mail was sent through http://webmail.uni-jena.de
Lapack mailing list
Julie Langou; Research Associate in Computer Science
Innovative Computing Laboratory;
University of Tennessee from Denver, Colorado ;-)
-------------- next part --------------
An HTML attachment was scrubbed...