ABSTRACT
The main idea of the multi-frontal solver algorithm working over the four finite element mesh
is illustrated in Figures 8.2, 8.3, 8.4, 8.5, and 8.6. In this example we assume that we also have
four processors (as many processors as elements). In such a case each element can be processed in
parallel . We will use the name processor, but core would be also suitable, depending on the assumed
architecture of the parallel machine. We also assume that we send some data from one processor to
the other. The communication is necessary in the case of a distributed memory parallel machine, but
it is not necessary in the case of a shared memory parallel machine.