LAPACK Archives

[Lapack] Bringing up lapack on a new architecture (origins of CLAPACK?)

I am pleased to report that the combination of gfortran (gcc 4.6.0) + dragonegg 
+ clang + a bit of python scripting to drive it all enables me to build all of 
lapack/lapacke (3.5.0) and successfully run the tests (correctly) in our 
instruction set simulator.

Thank you,

        Richard Gorton
        Cognitive Electronics
        rcgorton@Domain.Removed


On Sep 5, 2014, at 7:41 AM, Richard Gorton <rcgorton@Domain.Removed> wrote:

Thank you - I'll look into both LAPACKE and dragonegg.

Rick

On Sep 4, 2014, at 3:51 PM, julie langou <julie@Domain.Removed> wrote:

Hi Richard,
The latest f2c library is included in the latest CLAPACK package - CLAPACK 
3.2.1 - You can download it from http://www.netlib.org/clapack/

As you know CLAPACK is no longer maintained. We would recommend  tofind a 
way to link directly with the Fortran LAPACK or use the LAPACKE (C standard 
Interface).
Did you check http://dragonegg.llvm.org?
That could help you bringing LAPACK to your LLVM based compiler.

Hope it helps
Julie
On Sep 3, 2014, at 12:06 PM, Richard Gorton <rcgorton@Domain.Removed> wrote:


Hi, 


I'm bringing up our libraries & runtimes on a new processor architecture, 
and just started in on lapack.  As our compiler is LLVM based, there is no 
Fortran front end, so I turned to using the venerable f2c tool.  Which 
works fine until I hit constructs that post-date F77, such as 
CYCLE/EXIT/RECURSIVE/LEN_TRIM?

I've modified my copy of f2c to support some of these, but when I hit 
MAXLOC, I started looking around, and discovered that the CLAPACK content 
already had support for LEN_TRIM and MAXLOC in the f2clib support code.  
Which strongly implies that there exists an f2c version which already has 
implemented these intrinsics/new constructs.

Does anyone know if this really is/was the case, and if so, where I can 
find said version of f2c?  I do not seem to be able to find one (so far) 
which handles the above-listed features.

Thanks in advance,

    Richard Gorton
    Cognitive Electronics
_______________________________________________
Lapack mailing list
Lapack@Domain.Removed
http://lists.eecs.utk.edu/mailman/listinfo/lapack




<Prev in Thread] Current Thread [Next in Thread>


For additional information you may use the LAPACK/ScaLAPACK Forum.
Or one of the mailing lists, or