Sign up for our newsletter and get the latest HPC news and analysis.
Send me information from insideHPC:


InsideHPC Guide to HPC in Manufacturing

This is the first article in a series on the role of HPC in manufacturing. You can download the complete article series in PDF format from the insideHPC White Paper Library.

Introduction – Bringing HPC to Manufacturers, Large and Small

There are about 300,000 manufacturers in the United States, according to the National Center for Manufacturing Sciences (NCMS).  Over 95% are characterized as small to medium sized manufacturers (SMMs) with less than 2,000 employees.

Although small in size, when compared to global firms like General Electric, Ford and Procter & Gamble, the SMMs pack a lot of clout.  NCMS says that they account for more than twice the total employment of all the Tier One manufacturers combined and generate about three quarters of the total global R&D output.

In fact NCMS has coined a new term – “backbone manufacturers” – to identify this important segment of the economy.  In its recently released Voice of the Customer (VOC) survey, NCMS notes that “Without the backbone, the biggest manufacturers simply could not operate.”

NCMS is one of many organizations and institutions attempting to bring high performance computing (HPC) to the backbone manufacturers – especially modeling, simulation and analysis (MS&A).  Although HPC has made major inroads into the Tier One manufacturers, the technology has, with some notable exceptions, received a lukewarm reception from the backbone manufacturers.

Guide to HPC ManufacturingMajor stumbling blocks include the cost of HPC clusters, lack of trained IT personnel and employees conversant with such disciplines as computational fluid dynamics (CFD), crash analysis, and finite element analysis (FEA). Backbone manufacturers are highly risk averse and focused more on short term results over days and weeks – not the many months and even years that an HPC commitment implies. Moving from a workstation environment to an HPC cluster is a difficult transition. Backbone manufacturers (and smaller lines of business within the larger manufacturers) are not fully using the power of HPC to centralize and advance digital simulation and foster cooperation between engineering and IT.

From an HPC vendor’s perspective, this is a major underserved market that has yet to really experience the many benefits that HPC can bring.

Issued in 2014, an IDC vendor profile, “Dell and Engineering Services for Manufacturers” reports that manufacturers are accelerating the pace of new and more complex product introductions.  Many companies, IDC says, are struggling with costly development cycles, delays to market, poor product life cycle planning, and a shortage or misappropriation of highly skilled engineering resources.

The adoption of HPC by manufacturers, both large and small, can help alleviate these problems.

When it comes to HPC, the larger manufacturers are not only concerned about obtaining peak performance, but also optimal power, cooling, data center density, as well as the implementation of parallel file systems, virtualization and larger storage systems.

Backbone manufacturers have a different perspective.  Their concerns include becoming familiar with HPC and its requirements and making the business case for the technology. They want to know about the feasibility of creating an affordable HPC infrastructure that includes optimizing codes on the right architecture and integrating storage solutions into this environment. Navigating the complex and often costly landscape of ISV licensing is another major concern. There is also an underlying Return on Investment (ROI) challenge with Backbone manufacturers that large, global enterprises have come to terms with. At the larger enterprises, HPC is an established R&D tool. This may not be the case in many smaller manufacturers.

Over the next several weeks we’ll dive into the following topics:

If you prefer you can download the entire insideHPC Guide to HPC in Manufacturing for free courtesy of Dell and Intel. You can download this PDF now from the insideHPC White paper Library.

Resource Links: