HP OpenVMS Systems Documentation

The conjugated versions of this routine, BLAS1$VCDOTC, BLAS1$VZDOTC, and BLAS1$VWDOTC return the dot product of the conjugate of the first n-element vector with a second n-element vector, as follows:

Vectors x and y contain n elements that are accessed from arrays x and y by stepping incx and incy elements at a time. The vectors x and y can be rows or columns of a matrix. Both forward and backward indexing are permitted.

The routine name determines the data type you should specify for arguments x and y. Specify the same data type for these arguments.

Rounding in BLAS1$VxDOTx occurs in a different order than in a sequential evaluation of the dot product. The final result may differ from the result of a sequential evaluation.

Example

C
C To compute the dot product of two vectors, x and y,
C and return the result in DOTPR:
C
        INTEGER INCX,INCY
        REAL X(20),Y(20),DOTPR
        INCX = 1
        INCY = 1
        N = 20
        DOTPR = BLAS1$VSDOT(N,X,INCX,Y,INCY)

      

The Obtain the Euclidean Norm of a Vector routine obtains the Euclidean norm of an n-element vector x , expressed as follows:

Format

BLAS1$VSNRM2 n ,x ,incx

BLAS1$VDNRM2 n ,x ,incx

BLAS1$VGNRM2 n ,x ,incx

BLAS1$VSCNRM2 n ,x ,incx

BLAS1$VDZNRM2 n ,x ,incx

BLAS1$VGWNRM2 n ,x ,incx

Use BLAS1$VSNRM2 for single-precision real operations.
Use BLAS1$VDNRM2 for double-precision real (D-floating) operations.
Use BLAS1$VGNRM2 for double-precision real (G-floating) operations. Use BLAS1$VSCNRM2 for single-precision complex operations.
Use BLAS1$VDZNRM2 for double-precision complex (D-floating) operations.
Use BLAS1$VGWNRM2 for double-precision complex (G-floating) operations.

RETURNS

OpenVMS usage:	floating_point
type:	F_floating, D_floating, or G_floating real
access:	write only
mechanism:	by value

The function value, called e_norm, is the Euclidean norm of the vector x . The data type of the function value is a real number; for the BLAS1$VSCNRM2, BLAS1$VDZNRM2, and BLAS1$VGWNRM2 routines, the data type of the function value is the real data type corresponding to the complex argument data type.

Arguments

n

OpenVMS usage:	longword_signed
type:	longword integer (signed)
access:	read only
mechanism:	by reference

Number of elements in vector x to be processed. The n argument is the address of a signed longword integer containing the number of elements.

x

OpenVMS usage:	floating_point or complex_number
type:	F_floating, D_floating, G_floating real or F_floating, D_floating, G_floating complex
access:	read only
mechanism:	by reference, array reference

Array containing the elements to be accessed. All elements of array x are accessed only if the increment argument of x, called incx, is 1. The x argument is the address of a floating-point or floating-point complex number that is this array. This argument is an array of length at least:

where:

Specify the data type as follows:

If n is less than or equal to 0, then e_norm is 0.0.

incx

where:

If you specify a negative value for incx, it is interpreted as the absolute value of incx.

Description

BLAS1$VxNRM2 obtains the Euclidean norm of an n-element vector x , expressed as follows:

Vector x contains n elements that are accessed from array x by stepping incx elements at a time. The vector x is a real or complex single-precision or double-precision (D and G) n-element vector. The vector can be a row or a column of a matrix. Both forward and backward indexing are permitted.

The public-domain BLAS Level 1 xNRM2 routines require a positive value for incx. The Run-Time Library BLAS Level 1 routines interpret a negative value for incx as the absolute value of incx.

The algorithm does not provide a special case for incx = 0. Therefore, specifying 0 for incx has the effect of using vector operations to set e_norm as follows:

e_norm = n^0.5*|x[1]|

For BLAS1$VDNRM2, BLAS1$VGNRM2, BLAS1$VDZNRM2, and BLAS1$VGWNRM2 (the double-precision routines), the elements of the vector x are scaled to avoid intermediate overflow or underflow. BLAS1$VSNRM2 and BLAS1$VSCNRM2 (the single-precision routines) use a backup data type to avoid intermediate overflow or underflow.

Rounding in BLAS1$VxNRM2 occurs in a different order than in a sequential evaluation of the Euclidean norm. The final result may differ from the result of a sequential evaluation.

Example

C
C To obtain the Euclidean norm of the vector x:
C
        INTEGER INCX,N
        REAL X(20),E_NORM
        INCX = 1
        N = 20
        E_NORM = BLAS1$VSNRM2(N,X,INCX)

      

The Apply a Givens Plane Rotation routine applies a Givens plane rotation to a pair of n-element vectors x and y .

Format

BLAS1$VSROT n ,x ,incx ,y ,incy ,c ,s

BLAS1$VDROT n ,x ,incx ,y ,incy ,c ,s

BLAS1$VGROT n ,x ,incx ,y ,incy ,c ,s

BLAS1$VCSROT n ,x ,incx ,y ,incy ,c ,s

BLAS1$VZDROT n ,x ,incx ,y ,incy ,c ,s

BLAS1$VWGROT n ,x ,incx ,y ,incy ,c ,s

Use BLAS1$VSROT for single-precision real operations.
Use BLAS1$VDROT for double-precision real (D-floating) operations.
Use BLAS1$VGROT for double-precision real (G-floating) operations. Use BLAS1$VCSROT for single-precision complex operations.
Use BLAS1$VZDROT for double-precision complex (D-floating) operations.
Use BLAS1$VWGROT for double-precision complex (G-floating) operations.
BLAS1$VCSROT, BLAS1$VZDROT, and BLAS1$VWGROT are real rotations applied to a complex vector.

RETURNS

None.

Arguments

n

OpenVMS usage:	longword_signed
type:	longword integer (signed)
access:	read only
mechanism:	by reference

Number of elements in vectors x and y to be rotated. The n argument is the address of a signed longword integer containing the number of elements to be rotated. If n is less than or equal to 0, then x and y are unchanged.

x

OpenVMS usage:	floating_point or complex_number
type:	F_floating, D_floating, G_floating real or F_floating, D_floating, G_floating complex
access:	modify
mechanism:	by reference, array reference

Array containing the elements to be accessed. All elements of array x are accessed only if the increment argument of x, called incx, is 1. The x argument is the address of a floating-point or floating-point complex number that is this array. On entry, this argument is an array of length at least:

where:

Specify the data type as follows:

If n is less than or equal to 0, then x and y are unchanged. If c equals 1.0 and s equals 0, then x and y are unchanged. If any element of x shares a memory location with an element of y , then the results are unpredictable.

On exit, x contains the rotated vector x , as follows:

where:

incx

where:

If incx is less than 0, then x is referenced backward in array x; that is, x[i] is referenced in:

where:

y

where:

Specify the data type as follows:

If n is less than or equal to 0, then x and y are unchanged. If c equals 1.0 and s equals 0, then x and y are unchanged. If any element of x shares a memory location with an element of y , then the results are unpredictable.

On exit, y contains the rotated vector y , as follows:

where:

incy

where:

If incy is less than 0, then y is referenced backward in array y; that is, y[i] is referenced in:

where:

c

Specify the data type (which is always real) as follows:

s

Specify the data type (which can be either real or complex) as follows:

Description

BLAS1$VSROT, BLAS1$VDROT, and BLAS1$VGROT apply a real Givens plane rotation to a pair of real vectors. BLAS1$VCSROT, BLAS1$VZDROT, and BLAS1$VWGROT apply a real Givens plane rotation to a pair of complex vectors. The vectors x and y are real or complex single-precision or double-precision (D and G) vectors. The vectors can be rows or columns of a matrix. Both forward and backward indexing are permitted. The routine name determines the data type you should specify for arguments x and y. Specify the same data type for each of these arguments.

The Givens plane rotation is applied to n elements, where the elements to be rotated are contained in vectors x and y (i equals 1,2,...,n). These elements are accessed from arrays x and y by stepping incx and incy elements at a time. The cosine and sine of the angle of rotation are c and s, respectively. The arguments c and s are usually generated by the BLAS Level 1 routine BLAS1$VxROTG, using a=x and b=y :

The BLAS1$VxROT routines can be used to introduce zeros selectively into a matrix.

Example

C
C To rotate the first two rows of a matrix and zero
C out the element in the first column of the second row:
C
        INTEGER INCX,N
        REAL X(20,20),A,B,C,S
        INCX = 20
        N = 20
        A = X(1,1)
        B = X(2,1)
        CALL BLAS1$VSROTG(A,B,C,S)
        CALL BLAS1$VSROT(N,X,INCX,X(2,1),INCX,C,S)