SCIENCE

Netlist has announced that its HyperCloud memory module is accelerating the performance time of MSC.Software Finite Element Analysis (FEA) simulations.  Overcoming memory constraints due to the large number of floating point calculations and large models in every analysis or simulation project, HyperCloud is reducing simulation times for MSC.Software applications by 21 percent.

"Using HyperCloud with our Nastran FEA simulations enables us to eliminate the high levels of input/output wait times and run complex simulations, with large models, greater than 15 million degrees of freedom, faster," said Joe Griffin, development engineer of MSC.Software.  "This has led to an overall increase in productivity and the ability to conduct more simulations at a given time per server.  With HyperCloud, we also are able to conduct high-performance testing of critical parts and assemblies for products like automobiles and planes, and further design enhancements while decreasing design times."

Scientific modeling, financial analysis, and engineering design users can now launch multiple runs per day and run multiple jobs on a standard two processor server thereby increasing design productivity.  Traditionally, these workloads require complex multi-node clusters which lead to increased datacenter costs.

"MSC.Software's validation of the performance benefits that HyperCloud delivers to high-performance computing applications is helping engineers solve their most complex design problems," said Steve McClure, vice president of marketing at Netlist.  "We are excited to work alongside MSC.Software in removing the memory barriers our mutual customers encounter when modeling complex real world aerodynamic, mechanical impact, or heat transfer interactions."

Additionally, with HyperCloud, MSC.Software is now able to more accurately analyze structural responses to design factors, such as stress, strain, vibration, displacement and temperature.  Utilizing a patented architecture with advanced ASICs, HyperCloud allows memory to be populated at higher densities and speeds versus standard JEDEC RDIMMs.  As the first RDIMM compatible memory designed specifically to overcome the bottlenecks manifested in today's 64-bit processors, HyperCloud reduces cache misses that are common when simulating large models.