PRACE NPT

      Function dran0()                               Result( ran )

! -----------------------------------------------------------------

      Use        numerics

      Use        ran_module

      Implicit   None

! -----------------------------------------------------------------

! --- dran0 returns a uniform deviate in [0,1).

!

! --- The algorithm is loosely based on an algorithm from

!     Press & Teukolsky et.al. and based on the linear congruential

!     method with choices for M, A, and C that are given by

!     D. Knuth in "Semi-numerical algorithms".

! --- Input parameters:

!     Integer  - a1, c1, m1, a2, c2, m2. The parameters of the two

!                linear congruent relations used. They are passed

!                via module 'ran_module'.

!     Integer  - x1, x2. Seeds for the two linear congruences.

!                Passed via module 'ran_module'.

!

! --- Output-parameters:

!     Real(l_) - ran.  Uniform deviate in [0,1)

! ------------------------------------------------------------------

!

      Real(l_)            :: ran

! ------------------------------------------------------------------

      x1  = Mod( a1*x1 + c1, m1 )

      x2  = Mod( a2*x2 + c2, m2 )

      ran = ( Real( x1, l_ ) + Real( x2, l_ )*rm2 )*rm1

! -----------------------------------------------------------------

      End Function dran0