Notational Conventions
This document uses the following conventions:
Convention
| Description
|
|---|---|
lowercase courier | Indicates routine names.
|
lowercase courier italic | Indicates routine arguments.
|
The LAPACK routines are prefixed by a letter defining the data type:
Letter
| Description
|
|---|---|
s
| Indicates single precision: REAL, or REAL*4 in Fortran, float in C.
|
d
| Indicates double precision: DOUBLE PRECISION, or REAL*8 in Fortran, double in C.
|
c
| Indicates single precision complex: COMPLEX, or COMPLEX*8 in Fortran, single precision complex data type in C. See the note about complex data type below.
|
z
| Indicates double precision complex: DOUBLE COMPLEX, or COMPLEX*16 in Fortran, double precision complex data type in C. See the note about complex data type below.
|
i
| Indicates integer: INTEGER in Fortran, int in C.
|
?
| Indicates a place holder for a single letter in several routine names. For example,
?getrf refers to the routines
sgetrf ,
dgetrf ,
cgetrf , and
zgetrf .
|
Note
: Complex data type in C means a type containing two consecutive float or double fields representing real and imaginary parts of the complex number. Complex types
fcomplex
and
dcomplex
in oneMKL LAPACK examples are defined as follows:
/* Single precision complex datatype */ struct _fcomplex { float re, im; }; typedef struct _fcomplex fcomplex; /* Double precision complex datatype */ struct _dcomplex { double re, im; }; typedef struct _dcomplex dcomplex;