 |
|||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
|||||||||||||||||||||||||||||||||||
| Home >> Parallel Programming | |||||||||||||||||||||||||||||||||||
 |
|||||||||||||||||||||||||||||||||||
| Parallel Programming | |||||||||||||||||||||||||||||||||||
| Parallel programming techniques developed as a means of improving performance and efficiency. In a parallel program, the processing is divided into parts, with each part then being executed concurrently. The code from each part runs simultaneously on different CPUs. These CPUs can be on one machine or distributed over a network of CPUs. One of the first steps in building parallel programs is to identify sets of tasks that are able to run concurrently. If the same processing is required for a set of tasks with no dependencies in the computations and no communication is required between tasks, you have an ideal parallel computing opportunity. |
|||||||||||||||||||||||||||||||||||
| MapReduce Algorithm | |||||||||||||||||||||||||||||||||||
MapReduce is a distributed computing paradigm which allows you divide up tasks into smaller jobs and "map" them to a unique key. It is then possible to distribute these jobs over a compute grid and "reduce" all of the job results back into the original task The MapReduce algorithm allows for the simultaneous execution of software code and the processing of certain kinds of distributable problems using a large number of compute nodes. "Map" step: The master node takes the input, divides it up into smaller sub-problems and distributes those to worker nodes. "Reduce" step: The master node takes the results of all the sub-problems and combines them to get a final output, which
|
|||||||||||||||||||||||||||||||||||
 |
|||||||||||||||||||||||||||||||||||
The advantage of MapReduce is that it allows for distributed processing of the map and reduction operations. SEELE Compute Platform's implementation of MapReduce differs from other platform implementations by altering the cardinality or the type of mapping operation. In a conventional approach the worker nodes pull the sub-tasks for execution. SEELE Compute Platform's sub-tasks are pushed to the worker nodes and the process is initially controlled by the task. |
|||||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||||
