In the rapidly evolving AI landscape, GPUs have become indispensable, driving advancements from deep learning to complex data analytics. The latest AMD Instinct™ accelerators are purpose-built for exceptional performance and efficiency, delivering leadership capabilities across foundation model training, fine-tuning, and inference. However, AI is not a one-size-fits-all challenge—nor is enterprise IT infrastructure.

While GPUs remain essential for large-scale generative AI, 5^th Gen AMD EPYC™ CPUs are the world’s best CPU for enterprise AI[i] that provide a scalable, cost-effective path for AI-enhanced applications and host node performance. With high-frequency processing, leadership memory capacity, and seamless x86 compatibility, AMD delivers an evolutionary path to AI acceleration, enabling organizations to integrate, scale, and modernize their infrastructure at their own pace—with little disruption.

AMD EPYC Processors Scaling to New Heights at SC’24

Supercomputing 2024 has become a stage where technology leaders showcase innovations that push the boundaries of performance, efficiency, and scalability with announcements of the world’s largest superco

As the leader in silicon-to-systems design solutions, including electronic design automation (EDA) tools, an impressive portfolio of silicon-proven semiconductor intellectual property (IP), and pre-silicon hardware-assisted verification solutions, Synopsys is a key provider to drive semiconductor innovation.  Such contributions have broad impact to help the industry meet the unprecedented demands of high-performance computing, artificial intelligence (AI), and ever-upgrading devices and software.

In a recent AMD Tech Talk podcast hosted by AMD’s Jim Greene, Synopsys’ Senior Director of Product and Solutions Marketing, Todd Koelling, explained how Synopsys’ tools and expertise are enhanced with AI capabilities to empower semiconductor and systems companies to develop increasingly complex, yet energy-efficient AI processors, GPUs and accelerator chips. In this brief blog post, we’ll look at what Synopsys does and how AMD and Synopsys have collaborated in multiple areas to deliver state-of-the-art, high-complexity, energy-efficient chips.

Choosing the right infrastructure when deploying AI for business use cases is key. As the complexity of AI workloads increases, enterprises need to make informed decisions about memory footprint and compute capabilities, especially when dealing with multimodal AI that requires multiple models working together.  I recently had the pleasure of talking with Steen Graham of Metrum Ai about solving real-world AI challenges and the future of enterprise AI for the AMD EPYC TechTalk podcast series, available here. You can also find highlights from the interview below.  

Supermicro is one the world's premier producers of servers and storage systems, as well as a top provider of fast and efficient AI infrastructure. Linda Yang, a Supermicro senior solution manager, recently joined me on the AMD TechTalk podcast to discuss how the rapid evolution of AI models and workloads is driving change in AI infrastructure. I'm pleased to share highlights from the conversation in this blog post. You can listen to the full podcast here.

BeeKeeperAI operates a privacy-preserving, confidential computing platform that protects data and enables secure collaborations. At a time when security threats seem ubiquitous, BeeKeeperAI, with the help of security technologies built into AMD EPYC processors, help safeguard data. I had a great time speaking with Alan Czeszynski, BeeKeeperAI's marketing and product development leader. The full podcast can be accessed here.     

I always enjoy learning from people who understand how essential collaboration is to innovation. Alan Benjamin, CEO of GigaIO, is one of those people--and he was kind enough to spend some time speaking with us on our AMD TechTalk podcast. GigaIO, a company that provides workload-defined infrastructure for AI and technical computing, made news last year by configuring 32 AMD Instinct MI210 accelerators to a single-node server called the SuperNODE. Up to then, accessing 32 GPUs required four servers equipped with eight GPUs apiece — as well as all the cost and latency associated with linking up all of that extra hardware. I've included some of the interview highlights in the blog post below.

Our industry is rife with hyperbole and hard-to-defend and hard-to-replicate claims. AMD takes a different and more transparent approach. AMD EPYC servers make your Datacenter AI-Ready by improving performance and power efficiency without compromises, and with published numbers that speak for themselves.

AMD EPYC™ 7003 Series processors with AMD 3D V-Cache™ technology, formerly code-named “Milan-X,” were specifically engineered to excel at technical computing workloads. Today, AMD adds four of these processors to the EPYC 7003 lineup. Read this blog to learn more about the newest addition to the EPYC™ processor lineup.