Question

why is vectorization a faster alternative to loops?

Answer 1

Vectorization is a type of parallel processing. It enables more computer hardware to be devoted to performing the computation, so the computation is done faster.

Many numerical problems, especially solution of partial differential equations, require the same calculation to be performed for a large number of cells, elements or nodes. Vectorization performs the calculation for many cells/elements/nodes in parallel.

Vectorization uses special hardware. Unlike a multicore CPU, for which each of the parallel processing units is a fully functional CPU core, vector processing units can perform only simple operations, and all the units perform the same operation at the same time, operating on a sequence of data values (a vector) simultaneously.