Hi folks, i have a simple question regarding the GRASS BLAS/LAPACK interface. Why aren't we using the ATLAS library? http://math-atlas.sourceforge.net/
It provides a full BLAS implementations and some LAPACK routines. It has a native C interface, is tuned for many hardware platforms, is widely used and available for many operating systems. No need for a fortran to c converter, which was IIRC the major issue with the current gmath interface. The handling is not that complex as the current gmath wrapper, which handles the matrices in column order. ATLAS implements column and row major BLAS and LAPACK routines. Regarding the gmath library: The gmath code includes a G_matrix_free and a G_free_matrix routine, which is very confusing. And there are different ways of allocating memory for a **matrix and the mat_struct of the lapack wrapper. But the resulting matrix is the same, its a vector of the lengths col*row? I think we must simplify this. What do you think? Best regards Soeren -- Der GMX SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen! Ideal für Modem und ISDN: http://www.gmx.net/de/go/smartsurfer _______________________________________________ grass-dev mailing list [email protected] http://grass.itc.it/mailman/listinfo/grass-dev
