Paralelización de la reducción de una matriz simétrica a una matriz tridiagonal usando el lenguaje de programación ZPL

In this work the parallelization of the reduction from a symmetrical matrix to a tridiagonal matrix, using the ZPL parallel programming Language is shown. The description of the language ZPL is made, tests to matrices are made and the processing time, the acceleration and the efficiency are determined varying the quantity of processors. It is noted that for a fixed number of processors the processing time increases if the matrix size also increases, and for a fixed matrix size a clear relationship between the processing time and the quantity of processors doesn’t exist. Also it is noted that the largest percentage of processing time in each test is covered by matrix product operations.