{-------------------------------->  Vmath  <---------------------------------}
{ This unit contains vector and matrix procedures and functions for TURBO-   }
{ PASCAL, partly written as inline assembler code for a 387 coprocessor.     }
{ They are about two to three times faster than the equivalent "pure PASCAL" }
{ code.                                                                      }
{ Known features/limitations/bugs etc.:                                      }
{ - The unit has been written with TP6.0 on an 386SX/IIT387SX machine        }
{ - The procedure MulM4V4 needs an IIT coprocessor                           }
{ - The 287 coprocessor needs additional FWAIT commands in of strategic      }
{   places all over the code, since I don't have one I didn't bother.        }
{ - All routines PUSH DS on entry, use long pointers (You don't want to      }
{   be limited to 64K won't You ?) for operand access and POP DS on exit     }
{ - No testing of the routines has been carried out except that they work    }
{   fine and fast in my application - NO WARRANTY !                          }
{ - I wrote the routines as I needed them (or as I wanted to find out how to }
{   do it, in the case of MulM4V4) but at least the Vector3 operations are   }
{   quite complete by now. If I find the time some more Matrix3 code may     }
{   follow.                                                                  }
{----------------------------------------------------------------------------}
{ These routines contain no special artifice, but are straightforward        }
{ coded "mathematical common knowledge", so everybody is free to copy        }
{ and modify the whole unit or parts of it. And remember: Distributing       }
{ sourcecode advances the "Art of Computing" by allowing others to learn     }
{ from Your mistakes !                                                       }
{----------------------------------------------------------------------------}
{ I would be pleas