lapack_logical LAPACKE_dtp_nancheck	(	int	matrix_layout,
		char	uplo,
		char	diag,
		lapack_int	n,
		const double *	ap
	)

Definition at line 40 of file lapacke_dtp_nancheck.c.

{
    lapack_int i, len;
    lapack_logical colmaj, upper, unit;

    if( ap == NULL ) return (lapack_logical) 0;

    colmaj = ( matrix_layout == LAPACK_COL_MAJOR );
    upper  = LAPACKE_lsame( uplo, 'u' );
    unit   = LAPACKE_lsame( diag, 'u' );

    if( ( !colmaj && ( matrix_layout != LAPACK_ROW_MAJOR ) ) ||
        ( !upper  && !LAPACKE_lsame( uplo, 'l' ) ) ||
        ( !unit   && !LAPACKE_lsame( diag, 'n' ) ) ) {
        /* Just exit if any of input parameters are wrong */
        return (lapack_logical) 0;
    }

    if( unit ) {
        /* Unit case, diagonal should be excluded from the check for NaN. */

        /* Since col_major upper and row_major lower are equal,
         * and col_major lower and row_major upper are equals too -
         * using one code for equal cases. XOR( colmaj, upper )
         */
        if( ( colmaj || upper ) && !( colmaj && upper ) ) {
            for( i = 1; i < n; i++ )
                if( LAPACKE_d_nancheck( i, &ap[ ((size_t)i+1)*i/2 ], 1 ) )
                    return (lapack_logical) 1;
        } else {
            for( i = 0; i < n-1; i++ )
                if( LAPACKE_d_nancheck( n-i-1,
                    &ap[ (size_t)i+1 + i*((size_t)2*n-i+1)/2 ], 1 ) )
                    return (lapack_logical) 1;
        }
        return (lapack_logical) 0;
    } else {
        /* Non-unit case - just check whole array for NaNs. */
        len = n*(n+1)/2;
        return LAPACKE_d_nancheck( len, ap, 1 );
    }
}

Here is the call graph for this function:

Here is the caller graph for this function: