carlomat_4.0, a new version of the general purpose Monte Carlo program

A new version of the general purpose Monte Carlo program carlomat is presented that substantially improves efficiency of the phase space integration by automatic inclusion of parameterizations which map away the t-channel poles and peaks due to soft and collinear photon or gluon emission. The quadruple precision versions of the routines for computation of the helicity amplitudes and phase-space parameterizations have been written and calls to them implemented in the code generation part of the program. This allows to better control numerical stability of the Monte Carlo programs generated, in particular for reactions with a virtual photon or gluon exchange in the t-channel. A new option of generating s-channel kinematics has been added which takes into account peaks due to Feynman propagators of intermediate bosons which decay into on shell final state particles, or of the top quark. It speeds up both the compilation and execution time of the Monte Carlo programs with respect to the kinematics based on topologies of the Feynman diagrams used in versions 2 and 3 of carlomat. To further speed up the execution time, the main routine of the Monte Carlo program has been supplemented with the Message Passing Interface which allows to run the program parallelly on several processors. The previous version of this program (AEDQ_v3_0) may be found at https://doi.org/10.1016/j.cpc.2015.06.013.