[SciPy-user] Problems with ACML

Robert Kern robert.kern@gmail....
Tue Aug 28 13:12:33 CDT 2007


Ryan May wrote:
> Hi,
> 
> Does anyone here use the AMD Core Math Libraries (ACML) as their
> underlying libraries for BLAS/LAPACK/etc. in SciPy?  I have problems
> with (at least) scipy.linalg.eigvals (Fernando discovered this at SciPy
> on my laptop).  For instance, the following crashes reliably with ACML,
> but works fine with ATLAS versions of BLAS/LAPACK:
> 
>>> >from scipy.linalg import eigvals
>>> >from numpy.random import rand
>>>> a = rand(100,100)
>>>> eigvals(a)
> 
> Anyone else have this problem?  Is ACML known to be a bad thing to use
> with SciPy?

No, but it's not an option that's as thoroughly tested as ATLAS, either. If you
could supply a gdb backtrace from the crash, that would help locate the problem.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


More information about the SciPy-user mailing list