Don’t Just Upgrade Your Data Center—Transform Manufacturing Operations with AMD EPYC

Ryan Spurr
Ken Bartels

Operational excellence is everything in manufacturing, whether we’re talking about speeding research and development or producing next-generation products faster, cheaper, and at better quality. The companies that can bring products to market faster will outperform the competition.

Those manufacturers that value process and technology are not only investing in solutions to accelerate business processes but are also concerned with the ability to reduce bottom line costs, offset higher energy prices, meet sustainability objectives or investor obligations, and fuel innovation vital to delivering competitive advantage.

For example, 40% of carbon emissions come from the manufacturing industry and its consumption of electricity and fossil fuels to support operations. That same energy consumption impacts operational costs and runs counter to any corporate values, strategic objectives, or investor expectations for environment, sustainability, and governance (ESG) initiatives. Even for those organizations unconcerned about energy consumption, change is underway as regulatory organizations like the SEC issue proposed rulings that would mandate disclosure of greenhouse gases in financial statements. Whatever the driver, reducing energy consumption leads to reduced costs, and this is good business.

Moving away from energy to accelerating innovation, manufacturers shifting from legacy research, development, and production models to model-based product development achieve total savings between 70 and 80%. With the usage of digital twins, generative design and simulation, and integrated digital processes, manufacturers can dramatically reduce design cycles and costs—and bring products to market faster. Install new high-performance compute solutions such as these onto the highest performing and most efficient compute platforms, and your organization will see further benefits as compared to traditional solutions.

Productivity, another focused business measure in manufacturing, is also under assault as productivity fluctuates due to increased wage pressures, high employee turnover, material shortages, and higher costs. More recently, it was made clear how these headwinds were affecting businesses with Q1 2022 productivity dropping 7.5%, the most since 1947. With fluctuating productivity, inflation, and wage pressures impacting the bottom line, manufacturers look to “adopt new technologies or invest in equipment to make their workers more productive” and to attract and retain the talent necessary to sustain operations and grow their businesses.

Changes You Can Make Right Now

While many of these economic challenges are outside of manufacturers’ control, what actions can manufacturing leaders take to reduce costs, improve productivity, and also deliver on investor or regulatory initiatives, such as energy and sustainment?

One simple lever is investing in next-generation computing platforms. For example, AMD’s advancements in its compute platform are proving essential to optimizing how your business executes its strategic goals with realistic reductions in energy and infrastructure costs, and faster time to market.

By simply replacing the existing computing platform within your data center, research processes, or plant operations, you can obtain substantial business outcomes with a short time to value and an impactful return on investment.

Is your business looking to speed up research, improve processes, attract and retain the best talent, and out-compete the competition? Let’s look at how easy it is to invest in technology that has been proven to deliver results for manufacturers.

Business Levers You Can Pull

Manufacturers are under pressure to grow, cut costs, deliver faster, and continue to create more innovative people, processes, and tools to establish long-standing business differentiation in the marketplace. So, what are some of the typical business levers manufacturers might consider pulling to improve their technology posture while also delivering tangible business results in line with strategic objectives? AMD EPYC processors typically perform 17–50% better per-core than the market. This added performance is a significant factor in lowering data center operational costs in the following areas:

1. Rationalized Infrastructure: A common go-to for IT is to deliver the same or better services with a smaller footprint by rationalizing the number of servers, space, power, and associated costs to operate the data center. While the data center facility may not need to shrink, the demand on IT to deliver more compute surely exists, and having adequate space in the data center to house increased compute platform and workload is important for a growing business and deployment of new applications in support of the business.

With AMD EPYC, servers can be consolidated up to 8-to-1 over existing infrastructure. As an example, 50 servers can be consolidated down to six. This level of consolidation can reduce space in racks by 88%, rack floor space by 75%, administration costs by 88%, and power by 83%.

2. Reduced Licensing Costs: Perhaps more costly than the data center infrastructure itself is the software it’s designed to execute. Today’s ERP, PLM, and MES platforms are extremely expensive, and shifting to a lower density processor infrastructure without sacrificing compute capacity and performance may be one option to explore.

Software includes operating environment (OE) and applications licensing. The licensing models include per-processor and per-core costs but may also include other metrics like virtual memory. Deploying AMD EPYC can typically save 50–75% on per-processor OE licensing, including VMware virtualization software from vSphere to the full stack of VMware Cloud Foundation. For per-core OE licensing like Microsoft Windows Server, deploying AMD EPYC can save up to 50% over existing infrastructure.

The cost of applications, though, typically ranges from 5 to 20 times the cost of the computing hardware. And it is here where AMD EPYC makes a significant impact on lowering the cost of operations. For example, in a small 5-node database server cluster, switching to AMD with just 14% fewer cores needed to match the latest generation of Intel processors can save $272,000 on Microsoft SQL Server licensing, with an additional savings of $62,500 per year in support costs. The computing hardware in this example is around $75,000.

3. Increased Productivity: Nothing makes leadership more excited than increases in productivity measures, shortened design lifecycles, and reduced risk with new product introduction. With AMD’s compute platform, not only can you deliver reduced costs of operation, but you can also deliver more computing performance that results in more design cycles and simulations, as well as allow your organization to experiment with new design or production techniques before the product ever hits the factory floor. This not only creates a more productive organization, but also one that may be more innovative—and together those can deliver a significant market differentiation.

Productivity in IT operations is improved by having fewer servers or fewer processors to maintain. At any scale, deploying AMD EPYC improves reliability and availability and reduces security risk. Every EPYC processor has built-in Infinity Guard security features that work seamlessly without administrative burden.

Productivity in research, design, testing, and validation processes is improved when designers and engineers can do more in less time. Creativity and inspiration fuel innovation. Designers can do more simulations, raise element resolution to a design model, and add in more design factors in studies.

4. Reduced Energy Costs: Next-generation compute platforms are simply more efficient and utilize less energy. Whether the technology is housed in the data center servers, edge compute or edge servers in the factory, or in some specialty laboratory supporting engineering, AMD EPYC hardware will drive down operational costs, reduce peak time usage in at-risk locations, minimize utility penalties, and help contribute to any corporate objectives associated with environment and sustainability initiatives.

In the consolidation example shown above, the reduction of most interest here may be the reduction of greenhouse gases. For this example, emissions are reduced by 225 metric tons CO2 per year. Energy savings and emissions reduction associated with a consolidation with high-density compute in many cases fully justifies the cost of a replacement alone without any of the other factors considered.

Every manufacturer has different challenges and objectives, and AMD EPYC provides a full spectrum of processor core densities, from 8 to 64 cores per processor, that will enable a wide range of business stakeholders and use cases. Manufacturing applications will perform optimally at some point within this spectrum between deploying low-core density with higher per core performance and higher scale-out node count, or higher-core density with higher per node performance and fewer nodes. Throughout both ends of the spectrum, AMD EPYC excels over the competition in both performance and energy use. 

How Is Technology Applied to Business?

It’s important to not only evaluate the impact on IT (think space, maintenance, infrastructure costs), but also how investments will improve key business departments and the overall corporate business objectives. Here are a few use cases that Connection has supported for enterprise manufacturers and the kinds of outcomes achieved by updating to next-generation compute platforms.

1. High-performance Compute (HPC) Applications: Deploying AMD EPYC servers accelerates the design process for HPC applications, including computational fluid dynamics (CFD) and finite element analysis (FEA). EPYC provides higher model granularity and accuracy to the design process while speeding up the time to design and test the designs before prototypes are constructed and tested.

2. RTL Simulations for Semiconductor Design: Faster time-to-value (TTV) is realized with AMD EPYC, providing more simulations per day and faster staging of shared resources. RTL Simulation also provides design assurance, which reduces the risk of failure in a manufacturing step in a one-thousand-step process. A single failure can disrupt production, resulting in major costs and delays.

In both examples above, productivity gains become inherent as designers and engineers realize they can attempt new approaches and what-if scenarios. The other obvious benefits include shorter design and simulation cycles, more simulations in the same period, and increased productivity from your engineering teams.

3. ERP Infrastructure: AMD EPYC systems not only outperform in manufacturing design applications they also provide operational efficiency, reliability, and security in the data center. Enterprise resource planning (ERP) and other critical business systems are core to business operations. They can also act as a drag on business processes as software becomes more demanding, your business grows, and the amount of processing necessary to complete data processing or analysis tasks is constantly on the rise. This is especially true in common use cases such as material resource planning (MRP), netting, nightly processing, or closing books at the end of a financial period. Each of these tasks takes hours or perhaps days. These same software applications are also costly in terms of licensing and maintenance—creating even more opportunities to reduce overall licensing costs via hardware, while also improving overall performance and delivering results faster.

The Benefits of Server Consolidation

Let’s examine how AMD EPYC systems help to improve performance, reduce licensing costs, and deliver better reliability and security of critical business systems.

Operational efficiency is the result of deploying fewer servers, processors, and cores. Server consolidation is one method, and with EPYC, higher consolidation ratios are possible. Higher per-core capabilities maximize the number of users, improve response times, and speed up reporting and processing. EPYC per-core advantages average 23% over the competition, thus saving on per-core licensing. And saving that percentage on expensive software typically exceeds the cost of the server platform by a factor of 5 to 20 times. Other compute consumption models have similar cost savings.

Processor consolidation is another method that improves operational efficiency. It also improves reliability. Contrary to popular belief, reducing the processors in a server adds to reliability. Having 2-socket CPUs in a server doubles the risk of failure, as when one processor fails, it typically crashes the server and renders it unbootable until the processor is replaced.

Servers with four or more processors can be consolidated to one or two EPYC processors. A higher density of DRAM allows larger memory per CPU, so memory capacity is no longer an issue. The benefits include software licensing savings, lower energy, and smaller footprint.

EPYC provides better performance per watt which reduces power and cooling, a common issue with many data centers.

AMD EPYC’s Infinity Guard is a unique set of silicon-level security features built into every EPYC processor. These features are complementary to other data security solutions and work without additional software and the associated administrative burden. Infinity Guard includes secure memory encryption (SME), secure root-of-trust, and secure encrypted virtualization (SEV). Public cloud providers’ confidential computing is enabled by SEV. Infinity Guard protects from malicious attacks on physical memory access, malicious firmware attacks, and rogue attacks mitigating virtualization VM-to-VM and VM-to-hypervisor vulnerabilities.   

Start Reducing Costs and Improving Productivity Today

Deploying a modern compute platform is proven to improve performance and drive significant business outcomes in many areas of a manufacturing business. It’s easy to imagine how deploying AMD’s EPYC compute platform can benefit your business, especially if you have complex research and development, high-performance compute use cases, or even for ERP environments in search of better performance. Not only will AMD EPYC-enabled servers improve performance, but they can positively contribute to your space constraints, support energy reduction or sustainability initiatives, and most importantly, reduce operational costs across a range of business levers critical to IT and line of business. Don’t just upgrade—upgrade to the best, and transform how business gets done.

If your business is considering upgrading servers, or your company is interested in how next-generation compute platforms can improve how your business operates, then engage Connection’s AMD specialists and our Manufacturing Practice to learn more about how this technology and the many use cases may benefit your organization.

Ryan Spurr is the Director of Manufacturing Strategy at Connection with 20+ years of experience in manufacturing, information technology, and portfolio leadership. He leads the Connection Manufacturing Practice, go-to-market strategy, client engagement, and advisory services focusing on operational technology (OT) and information technology that make manufacturers more digitally excellent.