When Fast is Never Quick Enough
NX Nastran - the scalability and performance leader in complex system analysis
As the overall importance of CAE escalates within digital product development, there is a corresponding need for improved CAE performance. And, combined with seemingly inexhaustible advances in computer hardware, this trend is rapidly gaining traction in new product development and introduction, where fast is never quick enough.
Sophisticated digital simulations of extremely complex product performance can replace physical prototypes (for example, in areas such as auto crash, vehicle noise and vibration harshness, aircraft vibration and satellite systems thermal characteristics). The obvious benefit is cost, as each physical prototype can be extremely expensive. Virtual prototypes of this scale, however, require computer models of ever-increasing size and complexity. And taking that one step further, these larger model sizes call for more and more compute power to generate reliable results in a reasonable period of time - hours versus days or weeks.
Multi-computer environments consisting of numerous parallel processing nodes all connected via a high-speed network are becoming commonplace - especially where large-scale engineering computation requirements exist.
Intel's co-founder, Gordon E. Moore, has stated that, "...the first microprocessor only had 22 hundred transistors. We are looking at something a million times that complex in the next generation - a billion transistors. What that gives us in the way of flexibility to design products is phenomenal."
Figure 1 – Moore's Law Microprocessor Chart
Copyright © 2005 Intel Corporation
UGS is proactively taking advantage of these computing advancements by enabling its NX Nastran solution to be the scalability and performance leader in complex system analysis. NX Nastran can solve complex product performance simulations on up to 64 computers simultaneously. This can translate to major productivity improvements, given that other solutions in the industry are supporting only up to 16 simultaneous CPUs.
Image courtesy of DaimlerChrysler
In speaking about enhancements in NX Nastran, Paul Blelloch, Vice President of Aerospace Analysis at ATA Engineering, comments that, "I find the performance enhancements in distributive memory processing and the ability to handle very, very large models very exciting. We're extremely interested in new techniques to solve these models, and it's great to see that NX Nastran is on the leading edge in this area."
A UGS white paper discusses the fundamentals of these methods and highlights an industrial case study. Furthermore, as it is important to develop technologies that befit the current hardware architectures, the proper pairing of software technologies with appropriate hardware environments and future directions are presented in the paper, as well.
Learn how UGS is satisfying the perpetual need for increased digital simulation performance.
