Integrated grid computing comes to collaborative engineering users

It’s the first major fruit of the partnership deal signed 12 months ago between grid builder Platform Computing and Dassault. Platform has developed a version of its system integrated for Catia v5 r16, with costs ranging from $99 per CPU up to 1,000 per CPU, depending on the application. German automotive design and production giant Magna Steyr is among early adopters. By grid-enabling Catia, the firm says it was able to analyse the components of an entire vehicle in a single evening, as opposed to several days. “Grid technology from IBM and Platform Computing reduced the time required for our clash testing from 72 hours to four hours and contributed significantly to enhancing our design quality,” says Magna Steyr’s Dr Heinz Mayer. Thierry Ghenassia, Platform Computing alliances and partnerships director, says the development has been driven by customer demand for tools to improve and optimise CAD and PLM performance. “Platform LSF for V5 PLM is the first product to grid enable a PLM system. Our performance with LSF is very efficient and we also have a lot of experience in IT in the industrial manufacturing sector. Merging our products to optimise scheduling for a range of PLM applications within Dassault’s environment will provide significant improvements.” Dassault and Platform believe that as PLM expands to encompass much more of businesses’ inter-departmental processes, it becomes increasingly important for the IT infrastructure to manage compute intensive operations better. “The automation of all CAD, CAE, CAM and PDM operations to form a single multidisciplinary design optimisation loop will greatly streamline production but require thousands of CPUs to manage this new aggregate of operations,” they say. “Platform LSF for V5 PLM will drive resource optimisation and address the computing infrastructure challenges in the aerospace and automotive industries, resulting in reduced costs.” However, getting there requires some IT investment. As Chris Dudding, Platform systems engineer specialising in manufacturing applications says: “Users would have to buy a grid infrastructure, which could be a network of desktops and servers they already have. Then our LSF automatically controls the distribution of jobs to the grid, with batch workload management and resource allocation according to business rules.”