The Maple Grid Computing Toolbox enables you to run Maple computations in parallel, taking advantage of all your hardware resources, cutting down on processing time, and enabling applications that were not possible before.
The Maple Grid Computing Toolbox allows you to deploy your parallel programs to large-scale compute clusters and supercomputers, taking full advantage of all available processing power to tackle very large problems. This allows you to handle problems that are not tractable on a single machine because of memory limitations or because it would simply take too long.
The Grid Computing Toolbox is very easy to set up. It can connect directly into your existing Windows® HPC Server cluster without the need to set up services on each node. You can also start a server process on each computer on a network and the grid will self-assemble as each node automatically detects the other nodes that are present. The Grid Computing Toolbox also integrates into existing job scheduling systems such as PBS.
Maple includes the ability to easily set up multi-process computations on a single computer. The Maple Grid Computing Toolbox extends this power to multi-machine or cluster parallelism. The two versions are fully compatible, so that an algorithm can be created and fully tested using the local implementation inside Maple, and then deployed to the full cluster using the toolbox, without changes to the algorithm.
In order to perform distributed computations, the Grid Computing Toolbox offers a variety of tools:
- Standard MPI message passing on Windows HPC Server, for efficient communication and integration with tools that support this protocol.
- MPI message passing on Linux using the MPICH2 protocol, a high performance implementation of the MPI standard distributed by Argonne National Laboratory.
- An MPI-like message passing API, which is available on all platforms. This API is part of the Grid Computing Toolbox, so when using this protocol, no special drivers or other tools need to be installed on the computers in the grid. In addition to simplifying the setup of the grid, this protocol is also ideal for parallel computing on a single computer and computations across heterogeneous networks.
- A set of high-level commands in Maple for defining parallel computations.
- Distribute Maple computations across a network
- Integration with Microsoft Windows HPC Server, including Windows HPC Server 2008 R2
- Self-assembling grid in local networks
- Simple-to-use interactive interface for launching parallel jobs
- Tools to develop and test applications on your desktop before deployment
- Integration with PBS and other job scheduling systems
- MPI and MPI-like message passing API (send, receive, etc.)
- Automatic deadlock detection and recovery
- High-level parallelization commands (map, seq, etc.)
- Generic, parallel divide-and-conquer algorithm
- Many documented examples
- Support for heterogeneous networks
- Access to all the computational power of Maple
The Grid Computing Toolbox is available in two different versions:
Both versions are available from the Maplesoft Web Store.
- The Personal Edition supports up to 8 CPUs in the cluster.
- The Cluster Edition supports an unlimited number of CPUs in the cluster.
View system requirements
Note that a Maple license is required for each node in the cluster. Volume pricing is available for the node installations of Maple. Contact Maplesoft for details.