Prev Next RevMulvvOp

Reverse Mode Multiplication Operator

Syntax
inline void RevMulvvOp(size_t d,
     const 
Base *z, const Base *x, const Base *y,
     const 
Base *pzBase *pxBase *py)
inline void RevMulpvOp(size_t d,
     const 
Base *z, const Base *p, const Base *y,
     const 
Base *pzBase *py)
inline void RevMulvpOp(size_t d,
     const 
Base *z, const Base *x, const Base *p,
     const 
Base *pzBase *px)

Description
We are given the partial derivatives for a function  G(z, x, y) and we wish to compute the partial derivatives for the function  \[
     H(x, y) = G [ Z(x, y) , x , y ]
\]
where  Z(x, y) is defined as the d-th order Taylor coefficient row vector for Z as a function of the corresponding vectors for X and Y where
Operation Value
Mulvv  Z = X * Y
Mulpv  Z = P * Y
Mulvp  Z = X * P
Note that Z has been used both the original multiplication function and for the corresponding mapping of Taylor coefficients.

z
The vector z has length  d+1 and contains the d-th order Taylor coefficient row vector for Z.

x
The vector x has length  d+1 and contains the d-th order Taylor coefficient row vector for X.

y
The vector y has length  d+1 and contains the d-th order Taylor coefficient row vector for Y.

p
The scalar *p contains the value of the parameter P.

pz
The vector pz has length  d+1 and pz[j] contains the partial for G with respect to the j-th order Taylor coefficient for Z.

On Input

px
The vector px has length  d+1 and px[j] contains the partial for G with respect to the j-th order Taylor coefficient for X.

py
The vector py has length  d+1 and py[j] contains the partial for G with respect to the j-th order Taylor coefficient for Y.

On Output

px
If present, the vector px has length  d+1 and px[j] contains the partial for H with respect to the j-th order Taylor coefficient for X.

py
If present, the vector py has length  d+1 and py[j] contains the partial for H with respect to the j-th order Taylor coefficient for Y.
Input File: cppad/local/mul_op.hpp