LAPACK Archives

[Lapack] blas tests fail on x87

Dear Cezary,

The problem with the BLAS testing programs sometimes not computing the machine 
epsilon correctly has been fixed in revision r1224, available in the LAPACK SVN 
repository (


Rodney James
Assistant Research Professor
Mathematical & Statistical Sciences
University of Colorado Denver

On Mar 16, 2012, at 4:27 AM, Cezary ?liwa wrote:

BLAS testers does not determine correctly the machine precision and 
falsely reports errors. Details are given here:

This happens with GCC 4.6.x on x87.

Cezary Sliwa

Cezary Sliwa
Institute of Physics, Polish Academy of Sciences
al. Lotnikow 32/46, PL 02-668 Warszawa, POLAND

Lapack mailing list

<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