In 1964, The Control Data Corporation’s “CDC 6600” earned the illustrious title of “world’s fastest computer.” Three times faster than its closest competitor at the time (the “IBM 7030”), it performed up to 3 million instructions per second. It held onto that world-renowned recognition for four years, when it was surpassed by the company’s next-generation product, the “CDC 7600,” which could perform up to 30 million instructions per second.
Today, some 60 years later, high performance computing (HPC) capabilities process data and complex calculations at incredibly high speeds of quadrillions of calculations per second, a feat which mid-20th-century industry pioneers might never have imagined possible. HPC achieves these results with a technique called parallel processing, through which thousands of computers are networked together in a cluster that combines their power for greater speed and efficiency. Processing data at speeds measured in units called “floating-point operations per second,” or FLOPS, these super-computing systems also feature high-speed transportation of data between computer servers and ample storage to capture the output of all the data they manage.
The advent of high performance computing— which operates at speeds over a million times faster than the fastest commodity desktop, laptop, or server system in the market today— is enabling science, research, technology, business, artificial intelligence (AI), and society to advance in unprecedented ways.
Here are some of the many applications for high performance computing:
- Healthcare: Through activities such as molecular modeling and gene sequencing, high performance computing is used by medical researchers to predict how human cells will interact with specific drugs. This can improve diagnosis and treatment while also supporting the development of cures for diseases like cancer and diabetes. In one recent high-profile instance, HPC was effectively utilized by the medical community in conjunction with “The COVID-19 High-Performance Computing Consortium” (comprised of the Office of Science and Technology Policy, the U.S. Department of Energy, the National Science Foundation, and IBM) during the pandemic to quickly investigate the way COVID-19 cells invade and replicate within the body, supporting the rapid development of antiviral drugs to combat it.
- Energy/Climate Research: Through the creation of models involving massive amounts of historical meteorological and climate data, high performance computing currently supports everything from the forecasting of earthquakes, hurricanes, and other storms to wind simulation, climate modeling, insights on the formation of stars, the location and optimization of new oil wells, identification of new sources of renewable energy, and more.
- Media: The entertainment industry uses high performance computing to edit feature films, create amazing special effects, and stream live events worldwide.
- Finance: Among other financial services, high performance computing helps track real-time stock trends, automate trading, provide fraud protection, and identify cyber threats.
- Manufacturing: High performance computing helps manufacturers build products faster and more cost-effectively. In the semiconductor industry, for example, manufacturers are using the strength of HPC to increase productivity on the factory floor. Additionally, auto and engine manufacturers are currently using HPC to develop more fuel-efficient engines that could reduce fuel costs by over US$1 billion annually.
Entering the “Exascale”
With the ongoing digitization of business and society and the expanding availability of faster mobile connections, 3D imaging, and AI tools, experts predict that high performance computing will be in greater demand than ever by businesses across the spectrum as a way to address some of the world’s biggest challenges in science, engineering, and business.
According to industry professional Susheel Tadikonda, Vice President of Engineering in the Systems Design Group at Synopsys, an electronic design automation company, the field of high performance computing will continue to evolve based on the need for enhanced speed, memory, storage, and cloud security “to efficiently manage massive data volumes and deliver high processing and analytical capabilities to various sectors.”
This is already happening through the development of “exascale” computing. The term “exa” means 18 zeroes – specifically, exascale computing can perform more than 1,000,000,000,000,000,000 floating-point operations per second (FLOPS), or 1 exaFLOPS. In May 2022, the Oak Ridge National Laboratory’s Hewlett Packard Enterprise “Frontier” supercomputer officially became the world’s first and fastest exascale computer. It performed at a record 1.1 exaFLOPS, and experts believe it could evolve to perform at 2 exaFLOPS in the not-too-distant future.
Exploring the Future of High Performance Computing
Are you prepared for the Exascale Era and its capabilities?
Through High Performance Computing Technologies, Solutions to Exascale Systems, and Beyond, a five-course program from IEEE, learners will gain a better understanding of the history and evolution of HPC, from “big iron” computers decades ago to current and future exascale systems. Topics covered include the lessons history can teach about high performance computing, ways of accelerating application performance through hardware and software, the application of exascale computing to real-world problems, achievement of performance and efficiency in the HPC arena, and the use of AI and emerging technologies in science.
Connect with an IEEE Content Specialist today to learn how to get access to this program for your organization.
Interested in access for yourself? Visit the IEEE Learning Network (ILN).
Don’t Miss This Limited Time Special Offer!
In honor of IEEE Education Week, take advantage of this special offer on select eLearning courses— one of which is High Performance Computing: Use of AI and Emerging Technologies in Science— for US$10 each. Use the code ILNIEW24 at checkout on by 30 April 2024.
Resources
What is High-Performance Computing (HPC)? IBM.
BasuMallick, Chiradeep. (1 December 2022). What is a Supercomputer? Features, Importance, and Examples. Spiceworks.
Zhang, Kevin. (15 December 2021). High-Performance Computing has Become Crucial to Competitive Advantage—in Every Industry. Fortune.
Tadikonda, Susheel. Durrant, Scott. Knowlton, Scott. Molina, Ruben. (19 January 2022). Trends Driving the Future of High-Performance Computing (HPC). Embedded.
(22 November 2022). What is Exascale Computing? McKinsey & Company.
(4 February 2022). Researchers Tackle COVID-19 with AI. Caltech.
Platts, Leon. (10 November 2023). How the Semiconductor Industry is Leveraging High-Performance Computing to Drive Innovation. IBM.
(14 March 2022). The COVID-19 High-Performance Computing Consortium. National Institutes of Health.
What Is High Performance Computing? NetApp.
The finite element method (FEM) is essentially very complex math used by engineers to reduce the number of prototypes and virtual experiments necessary to create a successful design. In previous posts, we discussed the advantages of the finite element method (FEM) and finite element analysis (FEA). Together, FEM and FEA are used to predict the structural behavior and integrity of a design.
Specialized finite element analysis software emerged in the 1970s. Now, it is common to find virtual testing integrated into the product development cycle. The global simulation software market size reached US$11.08 billion in 2020. It’s expected to grow 17.5% by 2028 while the specific FEA software market is anticipated to grow nearly 9% over the same period.
Key factors to support this expected market revenue growth across several industries include the increasing need to reduce manufacturing costs, as well as the need to investigate critical situations without actual risks. Simulation software for problem solving and decision making will be important at almost every stage of manufacturing, including product design, testing, and market launch, to mitigate potential challenges and boost financial returns. These are just a few of the ways that various industries utilize FEM and FEA.
Manufacturing Industry
The manufacturing industry is facing problems due to a significant increase in manufacturing costs, rapid demand fluctuations, and excessive equipment investment. Consequently, the industry is faced with the challenge of simultaneously achieving eco-friendly, high-quality, and low-cost products. To meet these demands, organizations are making an effort to improve the efficiency of the manufacturing process using FEM to predict various variables such as die alignment, material size deviation, and working temperature.
Energy Industry
Currently, rethinking energy transport is essential due to its broader applications for different energy systems. The study of heat and mass transportation has received remarkable consideration by physicists, engineers, and mathematicians. Researchers are looking at how to boost thermal transportation by mixing the nanoparticles in the base fluid mixture. Utilizing FEA, they are working to find numerical and graphical outcomes related to velocity and temperature versus various parameters. The present developments are applicable in automobile coolants, as well as the dynamics in fuel and the production of solar energy.
Rail Industry
In designing for passenger rail vehicle safety, one of the most challenging tasks for design engineers in the rail industry is predicting material durability. Numerical simulation is a convenient solution for prediction challenges, but a model’s predictions strongly depend on the availability—and accuracy—of material and assembly data. Advanced adhesive properties can provide designers with a robust data package to address modeling challenges through complex calculation methods or FEA. For multiple passenger rail interior and exterior applications, 3M has characterized three of its structural adhesive technologies to meet data requirements for two safety classes.
Commercial FEA Software
Ansys Mechanical recently became one of the first commercial finite element analysis (FEA) programs supporting AMD Instinct™ accelerators, which are the newest data center graphics processing units (GPUs). The accelerators are designed to provide exceptional performance for data centers and supercomputers to help solve the world’s most complex problems.
“Today’s large, complex engineering challenges require quick, predictively accurate simulations that scale,” said Brad McCredie, corporate vice president at AMD. The collaboration between Ansys and AMD will enable a notable speed boost for applications, which will allow researchers to run complex structural simulations in order to drive higher quality, more efficient designs for cars, planes, and a range of other products.
Discover the Finite Element Method (FEM)
Learn one of the most powerful numerical approaches available to engineers. Finite Element Method for Photonics, a five-course program, covers the fundamental principles of FEM while providing participants with insight into the method.
Connect with an IEEE Content Specialist today to learn how to get access to this program for your organization.
Interested in access for yourself? Visit the IEEE Learning Network (ILN)
Resources
3M. (17 August 2022). Designing for passenger rail vehicles safety and durability. Railway Gazette International.
Ansys. (24 August 2022). Ansys and AMD Collaborate to Speed Simulation of Large Structural Mechanical Models Up to 6x Faster. Cision PR Newswire.
Brush, Kate. (Accessed 30 August 2022). DEFINITION: finite element analysis (FEA). TechTarget.
Emergen Research. (10 August 2022). Global Simulation Software Market Is Expected to Grow Steadily At CAGR Of 17.5% In The Forecast Period Of 2021-2028. EIN Newswires.
Ho Seo, Young. (2 August 2022). Development of smart cold forging die life cycle management system based on real-time forging load monitoring. Scientific Reports.
Infinium Global Research. (July 2022). Finite Element Analysis [FEA] Software Market: Global Industry Analysis, Trends, Market Size, and Forecasts up to 2028. Research and Markets.
Sohail, M., Nazir, U., El-Zahar, E.R. et al. (5 August 2022). Galerkin finite element analysis for the augmentation in thermal transport of ternary-hybrid nanoparticles by engaging non-Fourier’s law. Scientific Reports.
The finite element analysis (FEA) is leading to major breakthroughs in nanotechnology, and having a huge impact on a number of industries spanning electronics, material science, quantum science, engineering, and biotechnology, AZO Nano reported.
Simulations based on FEA, a complex mathematical technique, is giving engineers valuable insight into the mysterious mechanical properties of polymer nanocomposites used as filler in polymer manufacturing and processing. These properties offer a revolutionary alternative to conventional polymer composites, including enhanced abrasion resistance, less shrinkage, and residual stress, as well as advanced thermal, electrical, and optical properties.
Nanomaterials are much smaller than traditional materials, and are therefore typically not as effective. As such, it is crucial for engineers to understand how the materials will react under stress in order to improve their design. While FEA is just one technique used to test these designs, its unique abilities provide significant insight into their properties.
What is FEA?
As we previously reported, FEA is based on the finite element method (FEM), a technique that can help solve highly complex math equations. A simple way to understand FEM is to look at it as separating a large problem into a series of smaller ones (“finite elements”), making the overall problem easier to see. FEA is the mathematical equations behind FEM that is applied to create a simulation. The simulation breaks down the entire model into smaller elements within a mesh, which engineers use to test how the different elements of a design interact and perform under simulated stressors.
There are many benefits of using FEA. For one, its insight into how the various elements of a design are interacting in minute detail provide enhanced accuracy of structural analysis. Furthermore, FEA allows engineers to create virtual simulations thereby reducing the need for physical prototypes and testing in order to save time and money.
How Are Engineers Using FEA for Advancing Nanotechnology?
Using FEA, researchers have discovered that high interfacial stress can cause the nanofiber or matrix in the material to come apart. They were able to control the properties which improve the strength of the interface to generate the best stress transfer. They discovered that the accumulation of stress concentrations at the interface between the fiber and the matrix can reveal the effective matrix-to-nanofiber stress transfer. Additionally, engineers can use FEA to simulate the composition of nanocomposites and nanotubes in the polymer, which would strengthen their mechanical properties by organizing thousands of nanotubes in a specific pattern.
What Industries Are Benefiting From This Research?
The aerospace sector is using FEA to model and test the effectiveness of polymer nanocomposites-based structures. FEA is also used by the manufacturing sector to simulate the necessary properties of polymer nanocomposites for use in packaging and coating applications.
Engineers are also using FEA to make breakthroughs in the field of photonics. Examples include using FEA to analyze four-wave mixing of topological edge plasmons in graphene metasurfaces; to demonstrate a feasible way to control light on integrated photonics and free-space metasurfaces; and to develop advancements in surface-emitting semiconductor lasers and optical lenses. Learn more about how photonics researchers are using FEA to advance their field.
Problem-Solving Applications with Photonic Devices
Providing a comprehensive and up-to-date account of FEM in photonics devices, with an emphasis on practical, problem-solving applications and real-world examples, Finite Element Method for Photonics is a five course-program from IEEE. Created by Dr. Agrawal, learners will gain an understanding of how mathematical concepts translate to computer code finite element-based methods after completing this program.
Connect with an IEEE Content Specialist today to learn how to get access in order to train your organization.
Interested in the course program for yourself? Visit the IEEE Learning Network (ILN).
Resources
Ahsan, Muhammad Adeel (27 April 2022). Finite Element Analysis of Polymer Nanocomposites. AZO Nano.
Researchers from Tokyo University of Science have developed a new design method that could lead to lighter, faster, and cleaner vehicles and airplanes. Their technique, published in Composite Structures, simultaneously optimizes fiber thickness and orientation. As a result, it reduces the weight of reinforced plastic parts commonly used in aerospace, civil engineering, and sports equipment.
Traditionally, efforts have mostly focused on enhancing the strength of carbon fiber composites. However, the Tokyo researchers’ new design method optimizes both fiber thickness and orientation. Typically, carbon fibers are combined with other materials to make a composite, such as carbon fiber reinforced plastic (CFRP), which is popular for its strength, rigidity, and high strength-to-weight ratio. Some studies have examined how to improve CFRPs, particularly through a technique called “fiber-steered design,” which optimizes fiber orientation to enhance strength. The fiber-steered design approach, however, had a major flaw.
“Fiber-steered design only optimizes orientation and keeps the thickness of the fibers fixed, preventing full utilization of the mechanical properties of CFRP,” research team member Dr. Ryosuke Matsuzaki told Canadian Plastics. “A weight reduction approach, which allows optimization of fiber thickness as well, has been rarely considered.”
“Simultaneous Optimization Technique” Reduces CFRP Weight Without Affecting Strength
Faced with this dilemma, the researchers proposed a new design technique for simultaneously optimizing orientation and thickness depending on the composite structure’s location, which reduced the CFRP’s weight without affecting strength. According to their research, the method includes three phases.
- The Preparatory Phase:
During this phase, the researchers performed an analysis using the finite element method (FEM). As we discussed in a previous post, FEM is a numerical solution that breaks down a much larger, complex problem into a series of smaller ones (“finite elements”) in order to make the overall problem easier to examine. This equation is then used to create a digital simulation known as the finite element analysis, which gives engineers a more detailed look into the design and how its various elements work together. The team used the simulation “to determine the number of layers, enabling a qualitative weight evaluation by a linear lamination model and a fiber-steered design with a thickness variation model.” - The Iterative Phase:
The team implemented the iterative process to “to determine the fiber orientation by the principal stress direction and iteratively calculate the thickness using ‘maximum stress theory.’” - The Modification Phase:
During this step, the researchers made “modifications accounting for manufacturability by first creating a reference ‘base fiber bundle’ in a region requiring strength improvement and then determining the final orientation and thickness by arranging the fiber bundles such that they spread on both sides of the reference bundle.”
This simultaneous optimization technique led to a weight reduction of more than five percent and allowed for higher load transfer efficiency than what fiber orientation achieves by itself. In the future, the method could reduce the weight of CFRP parts that support greener transportation systems.
“Our design method goes beyond the conventional wisdom of composite design, making for lighter aircraft and automobiles, which can contribute to energy conservation and reduction of CO2 emissions,” Dr. Matsuzaki told Canadian Plastics.
FEM analysis is becoming an increasingly popular research tool, including in the field of photonics, where the method has contributed to a number of recent breakthroughs. Check out some of the latest innovations in optics research supported by this simulation tool.
Finite Element Method (FEM) for Photonics
Learn how FEM can be used to model and simulate photonic components/devices and analyze how they will behave in response to various outside influences. The Finite Element Method for Photonics course program provides a comprehensive and up-to-date account of FEM in photonics devices, with an emphasis on practical, problem-solving applications and real-world examples. Engineers will gain an understanding of how mathematical concepts translate to computer code finite element-based methods after completing this program.
Connect with an IEEE Content Specialist today to learn how to get access to this program for your organization.
Interested in the course for yourself? Visit the IEEE Learning Network (ILN).
Resources
Tokyo University of Science. (24 May 2021). New optimization approach helps design lighter carbon fiber composite materials. ScienceDaily.
Tokyo University of Science. (2 June 2021). Tokyo researchers hit on new design method to reduce weight in reinforced plastics. Canadian Plastics.