“The entire technology industry is transforming, and the consumer computing segment is no exception. To put it in perspective, the smartphones that we carry today are several million times faster than NASA’s combined compute power that launched the first Apollo mission.â€
The coming of age of Big Data and AI-powered data analytics as tools of Digital Transformation has forced organizations across the world to analyze vast amounts of data to shape future strategies. The need to store, segment, and analyze large datasets has, in turn, resulted in a massive demand for high-efficiency processors that can support AI and ML workloads as well as heavy-duty data processing. In an exclusive interview with Toolbox, Raghu Nambiar, Corporate Vice President of Data Center Ecosystems and Solutions at AMD, talks about how AMD is meeting the rising demand for higher computing power, what are the key priorities of computing leaders as we enter the new decade, and what will be the biggest trends in the high performance computing industry in the coming years.
He also talks about how AMD is maximizing the efficiency of the processing between CPU and GPU architectures to offer customers a unique value proposition for AI and ML workloads, and how the company ensures the security of its processors even as it continues to churn out more powerful, efficient, and future-ready chips.
Key Takeaways From This Tech Talk Interview:
- Top tips on how to maximize efficiency in computing
- A peek at the key priorities of computing leaders today
- Trends to follow in high performance computing for 2020 and beyond
- Ways to secure a data center using secure processors
Raghu, to set the stage, tell us about your career path so far and what your role at AMD entails.
I am the Corporate Vice President of Datacenter Ecosystems & Solutions at AMD. I was previously the CTO of the Cisco UCS business and played an instrumental role in accelerating the growth of the Cisco UCS to a top data center compute platform. I have spent my entire career working on software and hardware ecosystems for data centers, both in research and business.
I joined AMD in 2017 to help build a comprehensive software and hardware ecosystem for the AMD EPYCâ„¢ processor family. By doing this, we can provide our customers with an easy out-of-the-box experience when they deploy their applications on AMD powered servers and cloud instances. I also have a CTO hat for Software and Solutions at AMD, making sure we are aligning our technology roadmap with end customer workloads.
I have been in the industry for over two decades in various roles. One area that I wanted to mention are my leadership positions on industry standards committees for performance evaluation and leading academic conferences. I believe in open ecosystems and enabling fair competition. I thank my industry peers for electing me to chair the industry’s first standards committee for benchmarking big data systems, and the industry’s first standards committee for benchmarking the Internet of Things (IoT). I am the founding chair of TPC’s International Conference Series– a leading forum on Performance Evaluation and Benchmarking.
I have written more than 75 papers and book chapters, and in my book, Transforming Industry Through Data AnalyticsOpens a new window , I examined the role of analytics in enabling digital transformation, how the explosion in internet connections affects key industries, and how applied analytics will impact our future.
Learn more: CFO as Visionary: Using Digital Analytics to Power StrategyOpens a new window
In AMD’s vision for the future, high performance computing (HPC) takes the center seat. How does AMD plan to continue the required compute power growth, especially taking Moore’s Law into consideration?
We play in every major computing segment – from consumers, to enterprises, to HPC, to gaming and the cloud. Today, the HPC industry is transforming rapidly and touches every aspect of our lives. Some of the world’s most important and complex problems are being solved by HPC.
What you’ll see from AMD in the HPC industryOpens a new window is a continued effort for heterogeneous compute, which occurs when CPUs and accelerators work together to help increase the compute power and capacity of systems. Heterogeneous compute is one of the foundations of supporting the continued demand for more performance.
A great example of heterogeneous compute is the Frontier supercomputer that will be installed at the Oak Ridge National Labs in 2021. Frontier is projected to have an estimated 1.5 exaflops of performance when it goes live, using both AMD EPYC CPUs and AMD Radeon Instinctâ„¢ accelerators.
Beyond heterogeneous compute, the other areas where AMD is looking for ways to increase compute power is efficiency within the entire system. Whether that is support for faster memory, better interconnects or faster peripheral support, these are all things that will help streamline the compute power, so customers can do more, with less.
This is how AMD is advancing compute when Moore’s LawOpens a new window is slowing down. We are combining advancements in architecture, driving heterogeneous compute, looking at overall system efficiencies like right-sized core counts, large memory capacity, extreme memory bandwidth and massive I/O capabilities to enable exceptional HPC workload performance.
Learn more: How Relevant Is Moore’s Law and What Does It Means for the Tech Industry?Opens a new window
How will Big Data Analytics, Machine Learning, and Artificial Intelligence (AI) impact traditional HPC capacities?
The world is undergoing unprecedented change driven by technology advances that are creating enormous amounts of data in the process. This data represents an opportunity for companies to improve their business, create new revenue streams, and even invent whole new models to solve some of the world’s most challenging problems.
AMD enables companies to tap into the massive amount of data with powerful and effective processing power with the AMD EPYC processor and Radeon Instinct accelerators. In fact, AMD is the only providerOpens a new window in the world of both high performance x86 CPU and GPU technologies, and there is a role for both in artificial intelligence and machine learning. AMD Radeon Instinct GPUs are well suited for the training function of a deep learning model, while AMD EPYC CPUs are a great fit for inference.
Software is critical to both inference and training, as well as maximizing the efficiency of the processing between the CPU and GPU architectures. AMD is focused on large business opportunities where we can maximize the impact of both of our product families, in what we call “A+A†systems – AMD CPUs working with AMD GPUs. This allows us to apply our technical resources and software teams most effectively, giving our customers a unique value proposition for AI and ML workloads to use AMD supported capabilities like PCIe® 4.0.
In addition to hardware, AMD is committed to building an expansive software ecosystem with industry leaders. Enterprises of all sizes have realized the potential of Big Data AnalyticsOpens a new window (BDA) which is now commonly deployed on-premise, in the cloud, and in hybrid environmentsOpens a new window . We are focusing on technology partnerships, optimizations and easy-to-consume reference designs, both single-socket and dual-socket, that have our partners providing the flexibility, performance and scalability needed to help customers with their BDA.
Learn more: 60% Of European Tech Enthusiasts Prefer AMD Processors: EHA SurveyOpens a new window
What are the key priorities for consumer computing leaders today, given the focus on enabling inference models across companies and industries?
The entire technology industry is transforming, and the consumer computing segment is no exception. To put it in perspective, the smart phones that we carry today are several million times faster than NASA’s combined compute power that launched the first Apollo mission.
More and more things are getting connected, and in a data-driven economy speed is everything. AI and Big DataOpens a new window are now seemingly inseparable. The key priority for any computing leader today is identifying how to support inference in real-time at the data source or closer to it. This can be a consumer device or a specialized edge device.
At AMD, our mission is to provide the optimal compute technologies from edgeOpens a new window all the way to the core while offering unique security features to help protect your private data.
AI and ML security are big risk areas as CPU microarchitectures proliferate. What policies are needed to make CPU architecture safe for all stakeholders?
Security is top of mind at AMD. A key focus of the 2nd Gen AMD EPYC Processors are the security features to help defend applications and data. Data centers around the globe are constantly adapting to more securely meet the current workload demands while planning for future workload needs. To help fortify these data centers, the 2nd Gen EPYC uses two security features, Secure Memory Encryption (SME) and Secure Encrypted Virtualization (SEV). Additionally, the “Zen 2†architecture of the 2nd Gen EPYC is “hardened at the core†helping customers guard their most important assets, their data.
With SME and SEV, users can have greater confidence in the security capability surrounding their private data. The growing community of operating system vendors that support SEV includes Canonical, Fedora, Oracle, RedHat, and SUSE. VMware has also committed to support AMD security features in a future release of vSphere.
Learn more: Big Tech Forms Open-Source Partnership to Encrypt Cloud Data from Server to EdgeOpens a new window
What personal strategies do you use to work closely with functional leaders in order to understand what the business needs are for CPU architecture? What is the difference in the way a CIO and a CTO would approach these new architectures?
In the digital world, often a company’s differentiation is its ability to perform compute faster than its competition – this includes finance, retail, life sciences, oil and gas and many other industries.
We are a technology company, but our mission is to help solve problems efficiently and securely. Our value proposition is equally important to CIO and CTOs.
CIOs are responsible for running the business, focusing on business process, infrastructure, security and overall efficiency. CTOs are often excited about the latest and greatest and what’s coming up in the future. At AMD, we create products with both of their needs in mind.
With industry leading 7nm designs in CPUs and GPUs, we bring efficiency, high performance, density, and power efficiency to the data centers, helping lower the total cost of ownership (TCO) dramatically. The advanced security technologies we talked about earlier are designed to protect critical data about the business and their customers.
While there’s excitement about new technologies, we often see resistance due to uncertainties associated with new architectures and solutions. AMD provides close collaboration and joint engineering with our ecosystem partners to reduce the risks and give a seamless out-of-the-box experience when existing applications are migrated, and new applications are deployed.
Learn more: How AI Has Transformed the Role of a Chief Data Officer: LinkedIn CDOOpens a new window
Which trends are you tracking in this space for 2020 and beyond?
- Software Defined EverythingOpens a new window (SDx) – Virtualized environments are gaining more and more traction as users see the value and flexibility of intelligent software. We’re going to see an explosion of SDx from compute, storage, network, and new consumption models over the coming years.
- More migration to cloudOpens a new window – While many organizations still have on-premise needs for security and latency requirements, more and more data and apps continue to migrate to the cloud as providers offer specialized instances to meet unique workload demands.
- 5GOpens a new window – 5G will have a significant impact on economic growth. It will change the way we think about processing power at the edge. The increased data demands and coverage requirements for 5G nodes present exciting opportunities for both embedded processors and data center compute innovations.
- Edge computingOpens a new window – 5G has supercharged the race to make gains in compute power at the edge. Likewise, lower latency requirements for AI applications such as image recognition and neural language processing are driving many companies to ramp up their edge offerings.
- AIOpens a new window – One of the most exciting things on the horizon is the work we’re doing with the AMD powered Frontier supercomputer and the Center for Accelerated Application Readiness (CAAR)Opens a new window at Oak Ridge National Laboratory. These projects are applying machine learning in ways never before possible to help cure diseases, explore the Milky Way and develop new elements.
Learn more: Top 20 Technology Trends to Watch in 2020Opens a new window
About Raghu NambiarOpens a new window :
Raghu Nambiar is the Corporate Vice President of Datacenter Ecosystems & Solutions at AMD. He was previously the CTO of the Cisco UCS business and played an instrumental role in accelerating the growth of the Cisco UCS to a top data center compute platform. He has spent his entire career working on software and hardware ecosystems for data centers, both on in research and business use cases.
About AMDOpens a new window :
For 50 years AMD has driven innovation in high performance computing, graphics and visualization technologies ― the building blocks for gaming, immersive platforms and the data center. Hundreds of millions of consumers, leading Fortune 500 businesses and cutting-edge scientific research facilities around the world rely on AMD technology daily to improve how they live, work and play. AMD employees around the world are focused on building great products that push the boundaries of what is possible. For more information about how AMD is enabling today and inspiring tomorrow, visit the AMD websiteOpens a new window , blogOpens a new window , Facebook Opens a new window and TwitterOpens a new window pages.’
| About Tech TalkOpens a new window :
Tech Talk is a Toolbox Interview Series with notable CTOs and senior executives from around the world. Join us to share your insights and research on where technology and data are heading in the future. This interview series focuses on integrated solutions, research and best practices in the day-to-day work of the tech world. |
Would you like to share your thoughts about the evolution of the high-performance computing industry? Find us on TwitterOpens a new window , FacebookOpens a new window , and LinkedInOpens a new window . We’d love to hear from you!