Driver dndrv6

and .
The actions and are required. Note, the factorization of will generally need to be computed in complex arithmetic and stored as complex data.

• ido=-1
  Actions Required:

  Compute for .
  Set .
  The vector is in workd(ipntr(1)).
  The result vector must be returned in the array workd(ipntr(2)).

• ido=1
  Actions Required:

  Solve for .
  Set .
  The vector is in workd(ipntr(3)) (Action by ${\bf M}$ has already been made).
  The result vector must be returned in the array workd(ipntr(2)).

• ido=2
  Action Required:

  Compute .
  The vector is in workd(ipntr(1)).
  The result vector must be returned in the array workd(ipntr(2)).