LAPACK++ is a library for high performance linear algebra computations. The library includes support for solving linear systems using LU, Cholesky, QR matrix factorizations, and symmetric eigenvalue problems, for real-valued and complex matrices. In the current release, a bug with header files and a bug with uninitialized variables has been fixed. Also, the '++' in header file names has been removed, but wrapper files with the old name for backward-compatibility are still included.
The original LAPACK++ was released under a BSD-style license, but its development stopped in the year 2000 and the support for complex-valued matrices was not enabled anymore. With this new fork of the original LAPACK++, it is now possible to use a GPLed, full-featured LAPACK product in C++ again.