Re: numeric/ublas ?

From: walter_at_[hidden]
Date: 2001-11-12 16:40:13

--- In boost_at_y..., Toon Knapen <toon.knapen_at_s...> wrote:
> I was trying to build the boost/libs/numeric/ublas/tests but found 
out 
> they could not be compiled out-of-the-box due to peculiar 
dependencies 
> between the files (probably I did not get it how it should be done).
> But just to be sure, can someone tell me how they should be build 
best ?

ublas is somewhat scattered in the CVS (branch matrix_development). 
You'll find the odds at:
 
boost/boost/numeric/ublas    
Core library
 
boost/libs/numeric/ublas/bench1
Bench1 measures the abstraction penalty using certain dense matrix 
and vector 
operations.

boost/libs/numeric/ublas/bench2
Bench2 measures the performance of sparse matrix and vector 
operations.

boost/libs/numeric/ublas/bench3
Bench3 measures the performance of vector and matrix proxy's 
operations.

boost/libs/numeric/ublas/doc
Some documentation
 
boost/libs/numeric/ublas/test1
Test1 contains a couple of basic tests for dense vectors and matrices.

boost/libs/numeric/ublas/test2
Test2 demonstrates how to emulate BLAS.

boost/libs/numeric/ublas/test3
Test3 contains a couple of basic tests for sparse vectors and 
matrices.

boost/libs/numeric/ublas/test4
Test4 contains a couple of basic tests for banded matrices.

boost/libs/numeric/ublas/test5
Test5 contains a couple of basic tests for triangular matrices.

boost/libs/numeric/ublas/test6
Test6 contains a couple of basic tests for symmetric matrices.

We've never checked in any build related material, but we could add 
some makefiles (MSVC workspaces?) and build instructions.
 
> And what is the current status of ublas ? 
 
Rather advanced, w.r.t. Fortran BLAS functionality, I believe ;-). We 
are still writing docs and try to tune the kernels...
 
> Do you use Blas internally for 
> performance reasons (I checked but could no see any calls to the 
blas.h 
> header)
 
Nope.The template functions in blas.h are used to demonstrate how to 
emulate Fortran BLAS (see test2). However, if one needs platform 
specific tuned BLAS kernels, one could try to partially specialize 
these.

Regards
 
Joerg
 
> 
> toon