Industrial Applications of High-Performance Computing

253

Chapter 19

GE Research Industrial

Applications’ Journey

to Supercomputing

Rick Arthur and Masako Yamada

CONTENTS

19.1 General Electric’s Tradition for Innovation 254

19.2 GE and Computing 255

19.2.1 Shi to Modern State of the Art 255

19.2.2 Adoption of High-End Computing:

BarriersandSuccesses 257

19.2.3 e First “Ah-Ha!” Successes: Insights into

Portability and Scalability 261

19.2.4 Engaging the U.S. Government’s HPC Ecosystem 263

19.3 Case Study: Anti-Icing Surfaces 265

19.3.1 Project Background 265

19.3.2 Critical Role of the Greater HPC Community 267

19.3.3 Opportunity in Cold Climate Wind Energy Generation 268

19.3.4 Overview of Molecular Dynamics Techniques 270

19.3.5 Equally Important: Data Processing and Visualization 272

19.4 Molecular Dynamics Science 272

19.4.1 What’s Next 272

19.4.2 Industrial HPC Applications Landscape and Impact 273

19.4.3 Advanced Manufacturing, Big Data, and the

Industrial Internet 274

19.5 Conclusion 275

References 276

Endnotes 276

254 ◾ Industrial Applications of High-Performance Computing

19.1 GENERAL ELECTRIC’S TRADITION FOR INNOVATION

World-wide, the name omas Alva Edison is synonymous with

Invention. In founding the General Electric Company (hereaer “GE”) in

the 1800s, Edison created an institution for industrial innovation. Over

the past century, engineers and scientists at GE have shaped numerous

technological transformations in our lives. Starting with technology

to generate and transmit electricity to businesses and the home, that

electricity then powers GE lighting, motors, and appliances. GE pioneered

the technologies enabling the broadcast of radio and television to the

home. GE jet engine advances endowed military and civilian aviation

with speed, range, and eciency. GE technology leadership in diagnostic

imaging (computed tomography [CT] and magnetic resonance imaging

[MRI] scans) has nearly erased the phrase “exploratory surgery” from the

practice of medicine.

Now, 120 odd years later, GE operates in over 100 countries with a strong

and persistent presence spanning not only the heritage businesses but

now also complex industries like locomotives, water treatment, nancial

services, life sciences, and healthcare IT. is breadth of business areas,

each distinguished by a depth of technology, creates the microcosm in

which GE Global Research

1

practices science and engineering. GE Global

Research serves as the company’s hub for research & development, works

closely with the technologists in the businesses to innovate improvements

and entirely new products, oen leveraging experiences from one

industry to another (for example, using medical CT scanner technology

for inspecting parts for a jet engine). Likewise, essentially all products

and services apply numerical tools or integrate computing components,

aording opportunities for common investments to deliver diversied

impacts.

Another term synonymous, in fact eponymous, with GE’s founder

is Edisonian,

2

based on a perhaps misleading record of the method of

invention used in discoveries of materials needed in electrical lighting,

microphones, etc. However, the culture associated with systematic search

through empirical trials has indeed been a hallmark of GE’s research

and development. ere are two very sound reasons for this: many of

the relevant sciences lacked robust comprehensive numerical models to

replace physical study, and second, many of GE’s products are by their

nature critical to public infrastructure and/or human health and safety. As

such, the burden to justify costs and time to use numerical methodologies