Subject: Re: [ublas] LAPACK/BLAS bindings
From: Peter Man (plwm2_at_[hidden])
Date: 2010-06-07 05:25:48
Sorry, I'm using ATLAS 3.8.3.
Your answers make sense - how would you suggest working around the getri
On 07/06/2010 10:01, Thomas Klimpel wrote:
> Peter Man wrote:
>> I would be extremely grateful for a swift response!!
> A swift response is easy, as long as you don't expect me to quickly solve the reported issues.
> The errors you get from including<boost/numeric/bindings/lapack.hpp> are expected.
>>> What version of ATLAS are you using?
> You didn't answer this question.
> But I will not try again to reproduce it now, but only tonight.
>> However, if I then define BOOST_NUMERIC_BINDINGS_LAPACK_CLAPACK, then I
>> reproduce my runtime error.
> Well, this is expected, as only in this case the getri routine from atlas is used. As the problems seems to be that the workspace query returns a smaller than expected number, it is no surprise that the problem disappears as soon as you ask lapack instead of atlas for the size of the workspace.
> ublas mailing list
> Sent to: plwm2_at_[hidden]