Working with leading-edge computer platform and peripherals vendors to ensure our products are optimized on the widest variety of computer platforms.
Silicon Graphics International, Inc. (SGI) 46600 Landing Parkway Fremont, CA 94538 Country: United States Website:www.sgi.com
Business Contact: Tony DeVarco, Strategic Partner Manager 46600 Landing Parkway Fremont, CA 94538 Email: tdevarco(at)sgi.com Phone: 510-933-5124
Technical Contact: Scott Shaw, Applications Engineer 46600 Landing Parkway Fremont, CA 94538 Email: Shaw(at)sgi.com Phone: 734-564-3832
SGI® delivers a unified compute and storage solution to our manufacturing customers that reduces overall system management requirements and costs as well as simplifies data management and archival needs.
SGI's flexible x86 based server portfolio can scale effortlessly to meet our customer's compute and I/O requirements as they evolve. SGI® UV™ 2 product family provides the only single node, cache-coherent shared memory platform that can start small, and grow seamlessly as your needs develop. SGI UV 2000 is capable of consolidating the entire CAE workflow onto a single platform. SGI® ICE™ X and SGI® Rackable™ servers provide best of breed cluster computing. And finally, SGI® Modular Infinite Storage™ provides the ability to store and access the vast amount of engineering data created by these CAE bandwidth-intensive applications.
Solutions for Abaqus
Please download SGI Technology Guide for users of Abaqus FEA®. This guide is intended to help customers make knowledgeable choices in regards to selecting high-performance computing (HPC) hardware to optimally run Abaqus FEA software from the SIMULIA brand of Dassault Systèmes. This guide reviews the performance of Abaqus FEA executed on three types of SGI platforms: the SGI® Rackable™ cluster, the SGI® ICE™ X cluster and the SGI® UV™ 2000 Shared Memory Parallel (SMP) platform. In addition to presenting performance results on these three computer platforms, we discuss the benefits of using multicore Intel® processors, the trade-offs of different network topologies, NVIDIA® compute GPU device performance and the use of SGI® MPI PerfBoost™. Also included are sizing guidelines and recommendations for HPC computing platforms running Abaqus FEA.