From the course: Python Parallel and Concurrent Programming Part 2

Unlock the full course today

Join today to access over 22,600 courses taught by industry experts or purchase this course individually.

Agglomeration

Agglomeration - Python Tutorial

From the course: Python Parallel and Concurrent Programming Part 2

Start my 1-month free trial

Agglomeration

“

- In the first two stages of our parallel design process, we partitioned a problem into a set of separate tasks and established communication to provide those tasks with the data they needed. We looked at different ways to decompose the problem and focused on defining as many small tasks as possible. That approach helped us consider a wide range of opportunities for parallel execution. However, the solution it created is not very efficient, especially if there are way more tasks than there are processors on the target computer. - Now it's time to turn our thinking from abstract to something concrete and modify that design to execute more efficiently on a specific computer. In the third agglomeration stage, we'll revisit the decisions we made during the partitioning and communication stages to consider changes to make our program more efficient, combining some of those tasks and possibly replicating data or computations. As a parallel program executes, periods of time spent performing…

Contents