Star-P® Overview
Star-P software allows users performing scientific, engineering or analytical computation on array or matrix-based data to use parallel architectures such as multi-core workstations, multi-processor systems, distributed memory clusters and/or utility/cloud-based environments.
Today, Star-P delivers users of MATLAB® and NumPy (Numeric Python) the ability to:
- Transform serial applications for parallel deployment
- Develop and deploy new parallel applications and algorithms
The transformation or development of parallel applications is made possible by Star-P's "command set" and run-time system that extend these very high level languages (VHLL's) to the world of parallel systems. Star-P software addresses the key parallel programming challenges that face engineers, scientists and analysts (domain experts). This allows the domain expert to focus on their problem, algorithm and solution as opposed to low level parallel programming details.
- Productivity - development time
- Ease of use - abstraction of low level hardware details from development
- Performance - development time
- Deployable - reliable, accurate and usable by others
The Star-P software platform delivers revolutionary results to domain experts by enabling them to transparently use high performance computing resources, through the familiar desktop tools and languages they know best.
Data Parallel and Task Parallel Computing
Leveraging both data- and task-parallel computing is necessary in many scientific and technical simulations. Star-P enables users to work in both modes and to seamlessly interoperate between the two.
- Star-P's data-parallel mode enables algorithms requiring large-scale memory access and inter-processor communication, often called "global array computing", such as those found in matrix manipulation and signal processing applications.
- Star-P's task-parallel mode is ideally suited for parallelization of algorithms often called "embarrassingly parallel," where computations can be naturally broken up into largely independent processes such as Monte Carlo simulation, or parallelization of For loops.
Star-P allows users to quickly obtain significant performance speed up from server-based computing resources. For many users, application development is a significant portion of their solution time - a solution which requires extensive redevelopment of the application for parallelism would be counterproductive. This is why Star-P emphasizes easy-to-obtain parallel performance at reasonable levels of computational efficiency.
However, some applications are run repeatedly and obtaining high performance and efficiency can be crucial. For these cases, Star-P supports integration with C++ and FORTRAN as well as MPI-based kernels which provide the ability to tune to very high levels of performance, while still retaining the flexibility of the VHLL environment.
Learn More About Star-P
Star-P is a tool that can dramatically improve your productivity. We encourage you to learn more about Star-P by exploring the rest of our web site. The "Related Links" at the top of this page are a good place to start, or visit Getting Started with Star-P.