HP MLIB User's Guide Vol. 1 7th Ed.

234 HP MLIB User’s Guide
SGEMV/DGEMV/CGEMV/ZGEMV Matrix-vector multiply
Input trans Transposition option for A:
’N’ or ’n’ Compute y ←αAx + βy
’T’ or ’t’ Compute y ←αA
T
x + βy
’C’ or ’c’ Compute y ←αA*x + βy
where A
T
is the transpose of A and A* is the conjugate
transpose. In the real subprograms, ’C’ and ’c’ have the
same meaning as ’T’ and ’t’.
m Number of rows in matrix A, m ≥ 0. If m = 0, the
subprograms do not reference a, x, or y.
n Number of columns in matrix A, n ≥ 0. If n = 0, the
subprograms do not reference a, x, or y.
alpha The scalar α. If alpha = 0, the subprograms compute
y ←βy without referencing A or x.
a Array containing the m-by-n matrix A.
lda The leading dimension of array a as declared in the
calling program unit, with lda ≥ max(m,1).
x Array containing the vector x. The number of elements
of x and the value of lenx, the dimension of the array x,
depend on trans:
incx Increment for the array x, incx ≠ 0:
incx > 0 x is stored forward in array x; that is,
x
i
is stored in x((i−1)×incx+1).
incx < 0 x is stored backward in array x; that
is, if trans = ’N’ or ’n’, then x
i
is stored
in x((i−n)×incx+1); otherwise, x
i
is
stored in x((i−m)×incx+1).
Use incx = 1 if the vector x is stored contiguously in
array x, that is, if x
i
is stored in x(i). Refer to “BLAS
Indexing Conventions” in the introduction to
Chapter 2.
beta The scalar β.
’N’ or ’n’ x has n elements lenx = (n−1)×|incx|+1
otherwise x has m elements lenx = (m−1)×|incx|+1