ICL Research Profile

PLASMA

Overview

The Parallel Linear Algebra Software for Multicore Architectures (PLASMA) package is a dense linear algebra package at the forefront of multicore computing, designed to deliver the highest possible performance from a system with multiple sockets of multicore processors. PLASMA achieves this objective by combining state-of-the-art solutions in parallel algorithms, scheduling, and software engineering. Currently, PLASMA offers a collection of routines for solving linear systems of equations, least square problems, eigenvalue problems, and singular value problems.

PLASMA relies on runtime scheduling of parallel tasks, which is based on the idea of assigning work to cores based on the availability of data for processing at any given point in time. The concept, which is sometimes called data-driven scheduling, is closely related to the idea of expressing computation through a task graph, often referred to as the DAG (Directed Acyclic Graph), and the flexibility of exploring the DAG at runtime.

Find out more at https://bitbucket.org/icl/plasma

In Collaboration With

  1. University of California Berkeley
  2. University of Colorado Denver

With Support From

  1. Fujitsu
  2. MathWorks
  3. National Science Foundation
  4. The United States Department of Energy

Project Handout