Dear Desktop Engineering Reader:
The advantages of grid computing are well known: You can divide up an application’s processes and distribute them across a number of parallel CPUs so that you can efficiently run larger, more complex computations than ever before. The hassles of grid computing are also well known: To set it up right, you inefficiently expend an enormous amount of time pretending that you’re an IT type by assembling the grid, debugging your first 16 failures, and so on. Well, take heart Maple users. Maplesoft has rolled out its new Grid Computing Toolbox, and it’s the sort of thing that brings you one step closer to leveraging grid computing as an everyday thing.
The key to the Grid Computing Toolbox is simplicity. Setup involves starting a server process on each machine on a network. Your grid then self-assembles. The Grid Computing Toolbox also integrates with job schedulers like PBS, so you don’t have to do much fiddling with that stuff.
Even better, though, is what Maplesoft calls its “personal grid server.” This functionality lets you simulate a grid on your desktop workstation so that you can test and debug your parallel applications before you deploy them on the real grid. Also good news, the Grid Computing Toolbox lets you parallelize your large Maple-based numeric or symbolic computations for a multiprocessor workstation as well as Beowulf clusters and homogeneous or heterogeneous networks.
Still, the best part of the Grid Computing Toolbox is that it lets you be an engineer, not an administrator — no matter how much fun that might be — while enabling you to exploit the power of grid computing even if you live on a multiprocessor workstation. You can learn more about the Grid Computing Toolbox from today’s Pick of the Week write-up, where you’ll also find links to additional information from Maplesoft.
Thanks, Pal. – Lockwood
Anthony J. Lockwood
Editor at Large, Desktop Engineering magazine