[SciPy-dev] Linalg2 benchmarks
eric at scipy.org
Sun Apr 7 15:36:43 CDT 2002
----- Original Message -----
From: <pearu at scipy.org>
To: <scipy-dev at scipy.org>
Sent: Sunday, April 07, 2002 6:33 AM
Subject: Re: [SciPy-dev] Linalg2 benchmarks
> Hi Jochen,
> On 7 Apr 2002, Jochen Küpper wrote:
> > Well, knowing what Travis posted I must say what I find strange is t
> > he compared Numeric's lapack_lite with ATLAS where it is so easy to
> > use a machine optimized LAPACK/BLAS with Numeric. And btw. ATLAS
> > isn't always the best choice here -- one reason why I don't like this
> > tight binding to ATLAS too much.
> Using site.cfg it is possible to use your own lapack/blas
> libraries in favour of atlas ones by specifying a proper order.
> Though currently you still need atlas otherwise building cblas will fail.
> In future, we can remove the strict dependence of linalg on ATLAS easily
> as cblas or clapack routines are not used directly but through their
> wrappers blas.py and lapack.py.
Yes, I'm all for removing this restriction. Building SciPy on the large
parallel machines (O2K, SP3, etc.) his arder than it should be because ATLAS is
not always easy to build on these beast. However, they almost always have an
optimized lapack sitting around. Fixing this would probably make building on a
number of platforms easier.
This will take some cooperation between system_info and setup_linalg to test if
cblas actually exists.
> > something related: How much work would it be to get f2py/LinAlg to
> > work with numarray?
> Actually it should not be difficult. The most intensive use of Numeric
> array C/API is in fortranobject.c, basically, in the function
> array_from_pyobj and its dependencies. There is no need to change
> signature files. That's the whole beauty of using tools like f2py for
> generating extension modules automatically.
> When numarray will have a stable and documented C/API interface, I'll look
> for its support in f2py.
> > There is a numpy_compat now that supports almost all of Numeric, so
> > this would be one way. But in the long run one would like to have a
> > native implementation of linalg, of course...
> I am not sure that numpy_compat will work straightforward for f2py as f2py
> is quite aware of Numeric array internals. But on the other hand, I have
> not looked what is done in numpy_compat or numarray lately.
> Scipy-dev mailing list
> Scipy-dev at scipy.net
More information about the Scipy-dev